What We Have Learned about Terrorism since 9/11

Khusrav Gaibulloev Department of Economics

American University of Sharjah, UAE kgaibulloev@aus.edu

Todd Sandler* Department of Economics

School of Economic, Political & Policy Sciences University of Texas at Dallas Richardson, TX 75080 USA

tsandler@utdallas.edu

Journal of Economic Literature, June 2019, 57 (2), 275-328

This overview piece examines critically the post-9/11 empirical literature on terrorism. Major

contributions by both economists and political scientists are included. We focus on five main

themes: the changing nature of terrorism, the organization of terrorist groups, the effectiveness

of counterterrorism policies, modern drivers or causes of terrorism, and the economic

consequences of terrorism. In so doing, we investigate a host of questions that include: How

terrorist groups attract and retain members? What determines the survival of terrorist groups? Is

poverty a root cause of terrorism? What counterterrorism measures work best? In the latter

regard, we find that many counterterrorism policies have unintended negative consequences

owing to attack transference and terrorist backlash. This suggests the need for novel policies

such as service provision to counter some terrorist groups’ efforts to provide such services.

Despite terrorists’ concerted efforts to damage targeted countries’ economies, the empirical

literature shows that terrorism has had little or no consequences on economic growth or GDP

except in small terrorism-plagued countries. At the sectoral level, terrorism can adversely affect

tourism and foreign direct investment, but these effects are rather transient and create

transference of activities to other sectors, thus cushioning the consequences.

*While assuming full responsibilities for the paper, the authors gratefully acknowledge helpful

comments from Steven Durlauf and three anonymous referees on earlier versions.

What We Have Learned about Terrorism since 9/11

1. Introduction

The unprecedented four hijackings on September 11, 2001 (henceforth, 9/11) killed almost 3,000

people and injured over twice that number. These hijackings profoundly affected security

throughout the world and marked a watershed terrorist event. The deaths tied to 9/11 equaled

those from all transnational terrorist attacks from the start of 1988 through the end of 2000

(Sandler 2003). In addition, 9/11 gave rise to civil freedoms being traded away for security (see,

e.g., Davis and Silver 2004 and Viscusi and Zeckhauser 2003), large-scale (short-run) losses to

global stock exchanges, and significant reallocations of resources to counterterrorism measures

(Chen and Siems 2004 and Sandler, Arce, and Enders 2009). During October 2001, the Bush

administration launched a War on Terror that involved US allies and that primarily focused on

Africa, the Middle East, and Asia. These proactive measures, combined with actions to limit

terrorist financing, initially stressed al-Qaida and its loose network of affiliated groups (Sageman

2004, 2008). However, within a relatively short time frame, terrorist groups developed new

sources of funding (e.g., ransom payments from kidnappings)1 and exploited other means for

transferring funds (e.g., the Hawala system or money transfers through bank balances).

Subsequent terrorist bombings in the United Kingdom, Indonesia, Morocco, the Netherlands,

Saudi Arabia, Spain, Tunisia, Turkey, and elsewhere made the world acutely aware of the

mounting terrorist threat.

Following 9/11, the US government also channeled tens of billions of dollars into the

newly created Department of Homeland Security (DHS), whose budget increased by almost 9%

annually when averaged from 2003 to 2011 in nominal terms (Enders and Sandler 2012, pp. 328-

9). In more recent years, this increase is much more modest, (DHS 2016, p. 8). Other industrial

1 Brandt, George, and Sandler (2016) indicate some recent levels of these ransom payments.

2

countries also increased their security at home following 9/11, because their interests were also

targeted by the same religious fundamentalist terrorist organizations. The rise of al-Qaida in the

Maghreb (AQIM), al-Qaida in the Arabian Peninsula (AQAP), al-Shabaab, al-Qaida in Iraq,

Lashkar-e-Taiba, and the Islamic State of Iraq and Syria (ISIS) presents a dire threat to people

worldwide as these Islamic extremist groups exploit modern technology for propaganda purposes

and to radicalize followers. ISIS poses a particularly poignant threat as underscored by recent

high-profile attacks: suicide bombings, armed attacks, and truck rammings in Paris, Brussels,

Nice, Stockholm, London, and elsewhere.

With a few notable exceptions,2 economists took little research interest in applying

theoretical and empirical tools to study terrorism prior to 9/11. Most academic interest in

terrorism was displayed by political scientists, sociologists, and psychologists before 9/11.

These disciplines primarily contributed historical and descriptive studies on terrorism,

counterterrorism, and terrorist groups. After 9/11, the terrorism literature changed drastically.

Economists took a greater interest and applied econometric methods to analyze whether

terrorism has had significant consequences on macroeconomic aggregates, such as growth,

income, and investment. This is an essential issue because today’s terrorists are intent to damage

targeted countries’ economies, including key sectors such as tourism and foreign direct

investment (FDI) (Sageman 2008).3 Economists also examined the effectiveness of

counterterrorism policy and the root causes of terrorism.4 After 9/11, political scientists changed

2 A notable exception includes Landes (1978), who applies Gary Becker’s economic theory of crime to investigate the effectiveness of metal detectors, sky marshals, and other deterrents on US skyjackings, based on FAA data for 1961–1976. 3 Drakos and Kutan (2003), Enders, Sandler, and Parise (1992), and Sloboda (2003) are a few articles that examine the influence of terrorism on tourism. The relationship between terrorism and trade is investigated by Blomberg and Hess (2006) and Nitsch and Schumacher (2004), whereas the adverse impact of terrorism on FDI is studied by Abadie and Gardeazabal (2008) and Enders and Sandler (1996). 4For example, see Berman (2000, 2009), Berman and Laitin (2008), Enders, Hoover, and Sandler (2016), Krueger (2007, 2008), Krueger and Malečková (2003), and Kunreuther and Heal (2003).

3

their terrorism focus away from descriptive analyses to greater application of rational choice

models and econometric estimation.5 In the post-9/11 study of terrorism, the lines between

economics and political science became blurred, in which economists accounted for the

importance of institutions, politics, and context in their terrorism analyses, while political

scientists applied economic methods to their terrorism investigations.

Terrorism remains a very active field of research as terrorism continues to capture

headlines through its horrific attacks visited on the general public. The post-9/11 literature

contributes answers to the following questions: How has terrorism changed over the last half

century? How do terrorist groups attract and retain members? What determines the survival of

terrorist groups? What counterterrorism measures work best against terrorism? Should new

counterterrorism approaches be devised? Is poverty a root cause of terrorism? Have terrorists

had the damaging influence on targeted economies as intended? Many of these questions were

raised by the media or government leaders after 9/11 as a means of discovering a way forward to

address terrorism.

The primary purpose of this paper is to evaluate critically the post-9/11 empirical

literature on the political economy of terrorism. To keep our evaluation manageable and

focused, we consider five primary issues: (i) the evolving nature of terrorism, (ii) the role of

terrorist groups, (iii) the effectiveness of counterterrorism policies, (iv) some root causes of

terrorism, and (v) the economic consequences of terrorism. In so doing, we provide specific

answers to the questions raised above.

There are five general themes or insights that guide our analysis of these primary issues.

First, terrorism has altered in form after the rising dominance of religious fundamentalist 5For example, Bueno de Mesquita (2013), Carter (2010, 2015a, 2016), and Findley and Young (2012) explore the relationship between terrorism and insurgencies, while Benmelech, Berrebi, and Klor (2012), Dreher and Fischer (2010), Kurrild-Klitgaard, Justesen, and Klemmensen (2006), Li (2005), and Piazza (2008, 2011) investigate causes of terrorism.

4

terrorism in the 1990s and the augmented security measures in the West after 9/11. These

considerations have changed the lethality, location, and nature of terrorism over time, giving rise

to the need for new counterterrorism strategies. Second, terrorist groups respond rationally to

their environment to ensure their survivability and visibility. In so doing, they adopt novel

institutional forms and adjust their attack portfolios in response to counterterrorism actions.

Third, counterterrorism policies have had mixed success. Targeted governments often work at

cross-purposes relying too much on attack-deflecting defensive measures and too little on

proactive offensive measures, especially when the same terrorist group targets multiple

countries. Frequently, well-intentioned counterterrorism policies may have unintended

consequences as terrorists or governments strategically react to one another’s actions. More

thought needs to be given to countermeasures that offset terrorists’ actions, such as service

provision, that win them a constituency. Fourth, terrorism has myriad causes. The alleged

relationships between terrorism and globalization, terrorism and poverty, and terrorism and

regime type are much more nuanced than believed after 9/11. Fifth, as a general rule, terrorism

has had little direct negative impact on the economic growth or GDP of targeted industrial

countries despite some large-scale attacks.6 Any impact is felt by a few terrorism-fragile sectors,

and this impact is transitory and small relative to the economy. Larger macroeconomic effects

may plague small terrorism-ridden countries. Throughout our analysis, the underlying

theoretical framework is rational choice and strategic rational choice (or game theory).

The remainder of the paper contains seven sections. Section 2 offers some background

preliminaries, along with an empirical overview of terrorism since 1968. In Section 3, terrorist

groups’ goals, recruitment, longevity, organizational structure, and effectiveness are analyzed.

Section 4 presents some relevant game-theoretic notions that inform some post-9/11 empirical

6 There is, of course, an indirect effect on security expenditures, some of which may augment GDP.

5

studies of terrorism. The effectiveness of counterterrorism policies is taken up in Section 5,

followed by some select hypothesized causes and facilitators of terrorism in Section 6.

Macroeconomic and sectoral consequences of terrorism are analyzed in Section 7. Concluding

remarks are presented in Section 8.

2. The Evolving Threat of Terrorism since 1968

To address this issue, we must first be clear on a few definitions. Terrorism is the premeditated

use or threat to use violence against noncombatants by individuals or subnational groups to

obtain a political objective through the intimidation of a large audience beyond that of the

immediate victims (Enders and Sandler 2012, p. 4). This definition underlies the terrorist event

datasets, from which we draw our data for the subsequent graphs and tables. In the above

definition, there are two essential ingredients: the presence or threat of violence and a political

motive. Without such a motive, the violent act is a crime but not terrorism. Terrorist

perpetrators apply their violence in the hopes of cajoling a targeted government to concede to

their political demands. By making their acts appear to be random, the terrorists broaden their

audience way beyond the unfortunate victims to create an atmosphere of general anxiety. This

anxiety is meant to induce a country’s constituency (i.e., the audience) to apply pressures on the

government to concede to terrorist demands in order to restore safety. In other instances, the

terrorists seek a liberal democracy’s overreaction that undermines its legitimacy (Hoffman 2006

and Wilkinson 2001). Owing to the perpetrator’s subnational nature, standard terrorism

definitions rule out state terror, but do not rule out state-sponsored terrorism in which a state aids

a terrorist group with intelligence, logistical support, a safe haven, or resources.

Terrorism comes in two varieties that are germane to our study. Domestic terrorism is

homegrown and home directed, for which the victims and perpetrators are citizens of the venue

6

country, where the attack occurs. For instance, the bombing of the Alfred P. Murrah Federal

Building in Oklahoma City by Timothy McVeigh, an American, was an act of domestic

terrorism on April 19, 1995. Theodore Kaczynski’s, the Unabomber, package bombing

campaign on universities and airline companies during 1978–1995 was domestic terrorism. By

contrast, transnational terrorism involves two or more countries owing to the nationalities of the

victims and perpetrators, or the venue country sustaining the attack. If a terrorist incident

commences in one country but concludes in another, then it is a transnational terrorist event, as is

true for some hijackings, letter bombings, and assassinations. When the terrorists cross a border

to engage in their attack, the incident is transnational in nature. If one or more of the victims or

perpetrators are not from the venue country (e.g., the hijackings on 9/11), the terrorist attack is

classified as transnational. Transnational terrorist events often result in transnational

externalities as countries’ policies become interdependent. Increased defensive counterterrorism

measures, deployed by the United States and other targeted industrial countries after 9/11 (Bush

2002 and Enders and Sandler 2005a, 2005b), may displace planned terrorist attacks against these

countries to other less secure countries (see Sections 4 and 5). Additionally, US proactive

measures to weaken al-Qaida, an Islamic fundamentalist group engaged in transnational and

domestic terrorist attacks, limit the group’s effectiveness against all potential targeted countries,

thereby conferring a public benefit on such countries (Sandler 2005 and Sandler and Lapan

1988).

Prior to the 1990s, most terrorist groups were left wing or nationalist/separatist

(Blomberg, Engel, and Sawyer 2010, Cronin 2006, Hoffman 2006, and Rapoport 2004). The

rapid rise of the religious fundamentalist terrorist groups started in the 1990s with al-Qaida and

its Islamic extremist affiliated groups (Berman 2009, Gaibulloev and Sandler 2014, Phillips

2014, 2015, and Section 3). Unlike the leftists who generally wanted to limit casualties and

7

collateral damage, the religious fundamentalists wanted to maximize carnage as 9/11 and the

March 11, 2004 Madrid commuter bombings demonstrate. During the 1990s, the religious

fundamentalist assumed a dominant influence among terrorist groups.

[Table 1 near here]

Table 1 is based on Jones and Libicki’s (2008) list of 648 terrorist groups that started

before 2007.7 We list some characteristics of these groups before and after 2001 in the top and

bottom portions of the table. For the two periods, we distinguish groups by their ideology, e.g.,

left-wing indicates an underlying Marxist or socialist ideology, and nationalist denotes

nationalist/separatist groups. By resident country, we indicate two characteristics of the base

country: its income class and regime type. We also present the average annual number of

groups launched during 1968–2000 and 2001–2006. Using 2001 as the dividing line, we can

draw some interesting contrasts between left-wing and religious fundamentalist (denoted by

religious in Table 1) terrorist groups. First, religious fundamentalist terrorist groups represent

the greatest percentage of new groups started after 2001. This trend starts in the beginning of the

1990s; in our sample, 14% (44) of the groups formed before 1990 are religious fundamentalists,

whereas 30% (98) of the groups formed after 1990 are religious fundamentalists. On average, a

slightly higher annual number of leftist and nationalist/separatist groups are created during the

post-2001 period compared to the pre-2001 period. However, the average annual number of

newly-created religious fundamentalist terrorist groups more than tripled during the period after

2001. Second, prior to 2001, leftist groups are heavily represented in middle- and high-income

democracies, while the religious fundamentalist groups are heavily represented in low- and

middle-income autocracies and partly free democracies. Third, after 2001, these relative

distributions of leftist and religious fundamentalist groups remain essentially the same.

7 A more up-to-date list of terrorist groups that is matched to a terrorist event dataset is not currently available.

8

Because we do not have terrorist group data by the type of religion, we examine the

names of 52 religious fundamentalist groups that appeared during 2001–2006 and find that more

than 61% of those names contain words related to Islam. Out of 52 religious extremist groups,

23 emerged in Iraq, following the Iraq invasion, 3 began in Afghanistan, and 6 started in

Pakistan.

[Table 2 near here]

Table 2 displays regional distribution of the base of operations of 586 terrorist groups –

from Jones and Libicki’s list – that formed before 2002 and during 2002–2006. Groups are

distinguished by ideology. Some groups have multiple bases of operation in different regions, so

that the number of observations is greater than 586. We apply the World Bank regional

classifications: South Asia, SA; Europe and Central Asia, ECA; Middle East and North Africa,

MENA; Latin America and the Caribbean, LAC; East Asia and Pacific, EAP; sub-Saharan

Africa, SSA; and North America, NA. Overall, leftist terrorist groups are mostly concentrated in

ECA and LAC. Prior to 2002, nationalist terrorist groups appeared across all regions with large

numbers located in ECA, MENA, SA, and SSA. During 2002–2006, nationalist groups mostly

emerged in ECA and MENA. The religious fundamentalist groups started before 2002 located in

EAP, ECA, MENA, and SA. In the post-2001 period, the new religious fundamentalist groups

are based almost exclusively in MENA and SA. Comparing the base location of terrorist groups

that started before 2002 and during 2002–2006, the share of groups based in ECA and LAC

decreased in the post-2001 period, whereas the share of groups located in MENA and SA

increased in the same period. In particular, this share dropped from 31% to 26% in ECA, but

soared from 18% to 38% in MENA. Given that groups prefer to locate closer to the place of

their attacks (Gaibulloev 2015), this change is expected to raise the share of attacks in MENA.

Next, using the data from RAND (2012) and Jones and Libicki (2008), we track 103

9

terrorist groups that emerged after 2001 – see Table 3. By the end of 2006, almost 44% of the

leftist and nationalist/separatist terrorist groups ceased to exist. In contrast, 70% of the religious

fundamentalist groups remained active in 2007. Leftist terrorist groups in the sample account for

the greatest number of terrorist (domestic and transnational) attacks during 2002–2007.

However, attacks by the religious fundamentalist and nationalist/separatist terrorist groups are far

deadlier. The religious fundamentalist groups increased their activities and accounted for more

casualties (deaths or injuries) after 1993 (Berman and Laitin 2008 and Enders and Sandler 2000).

Unlike the leftists, some Islamic fundamentalist groups employed suicide bombings, which have

much higher death and injury tolls than conventional bombings.8 Sageman (2004, 2008)

attributes the increased lethality of religious fundamentalist terrorists, in part, to a “leaderless

jihad,” for which individuals can become radicalized and pose a homegrown threat that

traditional border protection is powerless to address. The Orlando nightclub massacre by an

American-born terrorist on June 12, 2016 is an example of this leaderless jihad.

[Table 3 and Figure 1 near here]

Figure 1 takes its data from two terrorist event datasets: International Terrorism:

Attributes of Terrorist Events (ITERATE) and the Global Terrorism Database (GTD).9 Both

datasets draw their observations from the news media. ITERATE only records transnational

terrorist incidents for 1968–2016; GTD records domestic and transnational terrorist incidents for

1970–2016. We use Enders, Sandler, and Gaibulloev’s (2011) method for splitting GTD into

domestic and transnational terrorist incidents. For all of our data displays, we rely on ITERATE

for transnational terrorist incidents and GTD for domestic terrorist incidents.

8 On suicide bombings, see Benmelech and Berrebi (2007), Pape (2003, 2005), Rosendorff and Sandler (2010), and Santifort-Jordan and Sandler (2014). A few nationalist/separatist groups – most notably, the Tamil Tigers and the Kurdish Workers Party – use suicide bombings. 9 ITERATE is maintained by Mickolus et al. (2017) and GTD is compiled by National Consortium for the Study of Terrorism and Responses to Terrorism (START) (2017).

10

In Figure 1, the number of transnational terrorist incidents is measured on the left-hand

vertical axis, while the number of domestic incidents is measured on the right-hand vertical axis.

The units on the right axis are quite compressed because domestic terrorist incidents outnumber

their transnational counterpart, depending on the time period. Transnational terrorism displays a

marked decline starting in 1994 as the religious fundamentalist groups grew in their dominance.

Following the end of the Cold War, some important European leftist terrorist groups

disappeared. Also the start of the 1990s witnessed much less state-sponsorship of terrorism

(Enders and Sandler 1999 and Hoffman 2006). In Figure 1, domestic terrorism far exceeds

transnational terrorism since 1977, with the exception of 1993. The one-year displayed decline

in domestic incidents in 1993 is the result of the GTD data for this year falling off of a truck

during a Pinkerton Global Intelligence Services (PGIS) office move.10 Since 2012, there is an

ominous rise in transnational terrorism due, in part, to many new Islamic terrorist groups, such as

ISIS, al-Shabaab, Boko Haram, and others increasing their campaigns of terror. Domestic

terrorism increased during 1975–1992 and 2008–2016. Some of the recent surge is due to more

generous coding of domestic incidents by GTD, especially since 2013 and a change in coders.

[Figures 2 and 3 near here]

Figure 2 presents the number of casualties per terrorist incident. Overall, the casualties

per attack increased for both domestic and transnational terrorism since the 1990s. Although,

there is a decline in the casualties after 2004, the numbers are still higher than those prior to

1990. The average numbers of casualties per transnational incident for the pre- and the post-

1990 periods are 2 and 8, respectively. The average number of casualties per domestic incident

is 3 before 1990 and 7 since 1990. The casualties per transnational terrorist incident continually

exceed the casualties per domestic terrorist incident after 2008.

10 PGIS originally collected GTD, which they gave to START (2017) to update and maintain.

11

In Figure 3, the proportions of transnational and domestic terrorist incidents with

casualties (i.e., deaths or injuries) are displayed. For the entire period, domestic terrorism is

associated with a greater proportion of incidents with casualties than transnational terrorism, but

this difference is closing since 2009. From 1999 on, there is a general upward trend in the

proportion of transnational terrorist incidents with casualties. This finding supports the notion

that the rising dominance of religious fundamentalist terrorism is associated with greater

carnage. Before 1990, on average, 26% of transnational terrorist incidents resulted in casualties,

while, from 1990 on, 46% of transnational terrorist incidents ended in casualties. This same

casualty percentage calculation for domestic terrorism gives 51% before 1990 and 67% for 1990

on. Since the early 1990s, the declining number of transnational terrorist attacks and the rising

carnage per attack strongly suggest that religious fundamentalist terrorists are getting more terror

from each of their attacks.

[Figure 4 near here]

Figure 4 displays the number and proportion of transnational terrorist attacks directed at

US people or property. One must remember that relatively few transnational terrorist attacks

occur on US soil, so that the bulk of these attacks are abroad. Since 2003, Figure 4 shows that

attacks against US interests are way down, given post-9/11 US defensive measures and US

proactive counterterrorism measures in the Middle East, Afghanistan, and the Philippines.

During 1968–2001, 32% of all transnational terrorist attacks were against US interests; during

2002–2016, 27.4% of all transnational terrorist attacks were against US interests. Just 15.3% of

transnational terrorist incidents were directed at US interests during 2013–2016. There is a

marked downward trend in this percentage after 2001, suggesting that counterterrorism measures

have kept US interests somewhat safer at home and abroad.

[Tables 4 and 5 near here]

12

Table 4 breaks down transnational terrorism and resulting casualties by its four major

modes of attacks – bombing, hostage taking, armed attack, and assassination. Bombings include

explosive, incendiary, letter, car, and suicide bombs, while hostage taking includes kidnapping,

skyjacking, barricade and hostage taking, and non-aerial hijacking. Other modes of attacks are

combined into one column under “Others.” As a general rule of thumb, bombings account for

about half of the total number of incidents and represent terrorists’ favorite type of attack. Most

kinds of bombings are logistically simple, requiring few resources. In Table 4, we distinguish

incident types for two periods: 1968–2001 and 2002–2016. In addition, we display the average

annual number of events, which dropped from 358 to 165, or by more than half, after 9/11, and

the average number of casualties per year, which surged from 957 to 1825. Table 5 provides the

same breakdown of incident types and time periods for domestic terrorist events. A noteworthy

feature is the increase in the average annual number of such incidents and casualties in the post-

9/11 period, indicative of a rising threat of domestic terrorism (Gaibulloev, Sandler, and Santifort

2012). For 1970–2001, domestic terrorist attacks outnumbered transnational terrorist attacks by

almost 4 to 1. This ratio is far greater in recent years. Similarly, the annual number of casualties

from domestic incidents greatly exceeds the annual number of transnational attack casualties.

[Figures 5 and 6 near here]

The changing patterns of terrorist modes of attacks before and after 9/11 are captured by

Figures 5 and 6, which display the percentage of transnational and domestic incidents by attack

type, respectively, for 1968–2001 and 2002–2016. For transnational incidents, there is a marked

drop in bombings and a pronounced increase in hostage taking (primarily kidnappings for

ransoms)11 and armed attacks after 9/11. Hostage taking rose as terrorists sought new sources of

11 See Brandt, George, and Sandler (2016), Brandt and Sandler (2009), and Gaibulloev and Sandler (2009a) on the composition of hostage-taking incidents.

13

funding as targeted governments went after terrorist-supporting charities and other means of

terrorist financing (Basile 2004, Bush 2002, and Enders and Sandler 2012). Armed attacks in

public places in major cities grew and, in so doing, allowed religious fundamentalist terrorists to

maximize carnage. A somewhat different picture emerges from domestic terrorist attacks for

these two periods in Figure 6. The proportion of bombings increased slightly after 9/11 and

remained well above 50%. Armed attacks also rose slightly and assassinations dropped as a

proportion of domestic terrorism. There was a noteworthy post-9/11 increase in hostage taking,

again indicative of terrorists seeking alternative funding sources.12

[Figures 7a and 7b near here]

Next, we turn to the pie charts in Figures 7a and 7b that show the changing regional

patterns of transnational terrorism attacks for the pre- and the post-9/11 periods. Following 9/11,

Figures 7a and 7b indicate that the share of transnational terrorist attacks decreased in ECA and

LAC and increased in MENA, SA, and SSA. This is due, in part, to the decline of the share of

leftist terrorists, who were heavily represented in ECA and LAC before 9/11, and the increase in

religious fundamentalist terrorists, who were heavily represented in MENA, SA, and SSA after

9/11. This switch in regional shares is also due to the increased security in NA and ECA after

9/11 and instabilities in MENA. We should also note that failed or weak states tend to be located

in the regions experiencing the post-9/11 increased proportion of transnational terrorist events.13

[Figures 8a and 8b near here]

A somewhat similar regional transition of the share of domestic terrorist attacks in the

pre- and post-9/11 periods is seen in Figures 8a and 8b. In the pre-9/11 era, domestic terrorism

was most concentrated in LAC and ECA; in the post-9/11 era, domestic terrorism was most 12 Some of the reason for the low pre-9/11 hostage-taking percentage is due to GTD undercounting these incidents – PGIS customers’ were foreign corporations that cared little about the abduction of domestic individuals. This undercounting changed once START took over the maintenance of the database. 13 On failed states and terrorism, see George (2018), Lai (2007), and Piazza (2008).

14

prevalent in MENA, SA, and SSA. Except for MENA, the post-9/11 percentages are different

between transnational and domestic terrorism. Many of the likely drivers for the regional

transition of domestic terrorism are the same as those for the regional transition of transnational

terrorism. One different driver concerns the location of civil wars, which can fuel domestic

terrorism (Findley and Young 2012).

We also construct similar figures for the regional distribution of attack casualties, where

the shifting patterns for casualties are similar to those of incidents. For transnational terrorism,

the share of casualties dropped markedly in ECA and LAC and rose in MENA, SA, and SSA for

the period after 9/11 – see Figures 9a and 9b. With regard to domestic terrorism – Figures 10a

and 10b – the share of casualties fell significantly in ECA and LAC and grew sharply in MENA

in the post-9/11 era, which is consistent with Figures 8a and 8b. However, the share of casualties

for domestic terrorist incidents remained constant between the two periods for SA and SSA, even

though both SA and SSA experienced an increase in the absolute number of casualties in the

post-9/11 period. The post-9/11 percentage declined in EAP. These are due to a dramatic

increase in the share of casualties in MENA resulting from recent conflicts and instabilities in the

region; MENA accounts for a half of all domestic incident casualties after the 9/11.

[Figures 9a, 9b, 10a, and 10b near here]

3. Terrorist Groups: Choices, Survival, Organization, and Effectiveness

A terrorist group is a subnational collective, consisting of members, who use terrorism to achieve

a political objective (Phillips 2015). Some terrorist groups – e.g., ISIS, Boko Haram, and

Fuerzas Armadas Revolucionarias de Colombia (FARC) – control territory, while most terrorist

groups – Italian Red Brigades, Direct Action, Black September, and Red Army Faction – do not.

In addition, some terrorist groups use a mix of terrorist and guerrilla attacks, in which the latter

15

are directed at military forces. Terrorist groups are often classified on the basis of four

ideologies: left wing, religious fundamentalist, nationalist/separatist, and right wing (Cronin

2006, Gaibulloev and Sandler 2013, 2014, and Jones and Libicki 2008). Issue-specific groups

are often labeled as left wing – e.g., groups promoting animal rights, groups seeking reduced

economic discrimination, or groups working for an end to the Vietnam War. If a particular

terrorist group adopts a Marxist orientation, then it is placed with the left-wing groups unless its

primary goal is the pursuit of national autonomy – Abu Sayyaf and Popular Front for the

Liberation of Palestine (PFLP) – in which case it is classified as nationalist/separatist. As shown

in Section 2, religious fundamentalist terrorist groups pose an increasing threat in terms of their

growing numbers and the lethality of their operations (see especially Table 3). Within these four

ideological-based terrorist groups, goals can differ greatly. Some groups aim for modest policy

changes, while others seek grandiose political transformations. In the latter case, Jemaah

Islamiyah campaigns for a Pan-Islamic state in Southeast Asia. Other terrorist groups want to

maintain the status quo over a proposed policy change. In numerous instances, terrorist groups

pursue territorial change, especially among nationalist/separatist organizations.

3.1 Terrorism, Insurgency, and Civil War

Prior to analyzing terrorist choices, we must distinguish among three concepts of political

violence. An insurrection is a “politically based uprising intended to overthrow the established

system of governance and to bring about a redistribution of income” (Sandler and Hartley 1995,

p. 307). Unlike terrorism, insurrections involve a large sector of the population that throws its

support behind rebel leaders. By contrast, guerrilla warfare is a tactic that relies on mobile rebel

forces to engage in irregular warfare against larger governmental forces. Like guerrilla warfare,

terrorism is a tactic that may or may not be employed in an insurrection.

16

Civil wars, also known as intrastate wars, consist of an armed conflict between a

sovereign state and organized domestic groups that engage a government’s military forces

(Sambanis 2008). Civil wars must surpass some threshold of deaths, which is not true for

terrorism. Findley and Young (2012) highlight a few specific instances wherein terrorist tactics

were deployed before, during, or after civil wars.14 Although terrorism may be used by rebel

forces, terrorism need not be associated with civil war and most frequently is applied in the

absence of such war – e.g., Italian Red Brigade, 17 November, Euskadi ta Askatasuma (ETA),

and Combatant Communist Cells (CCC) sought political goals without generating or being

involved in a civil war. During civil wars, rebel attacks against government forces are not

counted in terrorist event datasets because attacks against combatants are excluded. When these

attacks are directed against civilians or passive military targets (e.g., the 1983 suicide bombing

of the US Marine barracks in Lebanon) and are intended to influence an audience, the attacks are

considered to be terrorism. Most of the bombings in Afghanistan and Iraq following 9/11 are not

included in the standard event datasets. For 1970–2012, Gaibulloev, Piazza, and Sandler (2017)

indicate that 70% of all terrorist attacks occurred in countries that did not experience a civil

conflict.

Thus, there is a strong practice in the terrorism literature to treat terrorism, insurgencies,

and civil wars as distinct concepts or in isolation. However, the rise of ISIS, Boko Haram, al-

Shabaab, and AQIM, which control territory, means that the choice between guerrilla warfare

and terrorism tactics is relevant for a select number of terrorist groups.15 Consequently, we next

consider this choice along with other more commonly studied terrorist choices.

14 Also see Berman, Shapiro, and Felter (2011), Bueno de Mesquita (2013), Carter (2010, 2015a, 2015b), and Kalyvas and Balcells (2010). 15 According to Jones and Libicki (2008, p. 98) most terrorist groups are not engaged in an insurgency.

17

3.2 Terrorist Group Choices

Given the rise of some territory-controlling terrorist groups, there is increased theoretical and

empirical interest in explaining their choice between conventional war and irregular tactics –

guerrilla warfare and terrorism.16 If this choice is endogenous, then empirical models that treat

civil wars as an independent variable when explaining the determinants of terrorism are likely

misspecified. Bueno de Mesquita (2013) puts forward a two-stage game of this choice involving

a rebel group and a potential population of supporters, where the government is passive and

outside options are exogenous. Moreover, rebels must choose between conventional conflict and

irregular tactics – they cannot engage in both at a given time. Among other interesting results,

he shows that “the use of irregular tactics is highest in societies in which nonviolent opportunity

is at moderate levels, such that mobilization is low, but extremists are still willing to fight”

(Bueno de Mesquita 2013, p. 234). Thus, the use of terrorism and guerrilla tactics may not

necessarily increase monotonically as nonviolent opportunities decline, contrary to conventional

wisdom. We will see in various other places the prevalence of nonlinearities associated with

terrorist behavior.

In an alternative, but somewhat similar exercise, Carter (2016) puts forward a game-

theoretic model in which terrorists choose between guerrilla attacks and terrorism in an effort to

provoke a violent response from the government.17 In the model, the terrorist group goes first

and decides the type of attack, followed by the government’s choice of the forcefulness of its

response. An implication of his model is that the government’s response is likely to be stringent

if it can inflict greater damage on group members than on civilians. The terrorist group’s choice

of attack type depends on how provocative it intends the attack to be. Generally, guerrilla 16 Works on this topic include Berman, Shapiro, and Felter (2011), Carter (2015a, 2015b, 2016), de la Calle and Sánchez Cuenca (2012), Kalyvas and Balcells (2010), and Laitin and Shapiro (2008). 17 On terrorists provoking a violent response, see Arce and Sandler (2007, 2010), Carter (2012), Jacobson and Kaplan (2007), Jaeger and Paserman (2008), Kydd and Walter (2002), and Merari (1993).

18

attacks are anticipated to cause more damages, thereby evoking a greater state response. When

confronting weak states, terrorist groups are more inclined to choose guerrilla tactics because the

concomitant state’s forceful response is less feared. To test his hypotheses, Carter (2016) draws

his data from Terrorism in Western Europe Event Data (TWEED) (Engene 2007). Carter

classifies attacks against passive military targets (e.g., National Guard) and police as guerrilla

warfare and classifies attacks against other targets as terrorism. During the 1950–2004 period of

his study, almost all Western European terrorist groups, with the exception of the Irish

Republican Army (IRA), did not control territory. Moreover, attacks directed at passive military

targets and police are not instances of civil wars or guerrilla warfare in the sense discussed in the

theoretical literature. Since most terrorist event datasets exclude attacks against combatants,

there is a paucity of data that can test the endogenous choice between terrorism and civil wars.

Also, as stated previously, most terrorist groups do not control territory or engage in civil wars.

Another terrorist choice involves resource allocation between legal activities (e.g.,

running for office or engaging in protests) and terrorist attacks. In the literature, this choice is

examined with standard choice-theoretic models (Anderton and Carter 2005, Enders and Sandler

2004, and Frey and Luechinger 2003).18 These articles focus on substitution and income effects

associated with governments’ action to alter the prices of legitimate protest or illegitimate

terrorist attacks. The result of such government-engineered price changes on protests and

terrorism may be ambiguous for normal goods as substitution and income effects oppose one

another. To date, this choice is not studied empirically, primarily because there is no dataset that

tracks terrorist groups’ legitimate activities. The Jones and Libicki (2008) data indicate which

terrorist groups gave up terrorism and joined the political process (see the next subsection), but

this is an all-or-none choice. Terrorist groups’ choice between legitimate and illegitimate

18 Other choice-theoretic articles on this topic include Sandler (2010) and Siqueira and Sandler (2010).

19

activities is certainly worth studying and would limit some of the underlying exogeneity of many

empirical investigations of terrorist groups’ behavior.

A more studied choice of terrorist groups involves their campaign composition among

alternative modes of attacks in reaction to counterterrorism measures. Enders and Sandler

(1993) investigate this choice in terms of a characteristics model of consumption in which

terrorist modes of attack that provide the same characteristics (e.g., ransoms) are substitutes.

When attack modes must be combined to produce a basic commodity or characteristic (e.g., an

atmosphere of public anxiety), these modes are complements. Among other results, these

authors argue that airport metal detectors should result in fewer hijackings but more kidnappings.

Given that these detectors were later deployed in embassies, these detectors are also expected to

reduce terrorist attacks on embassy grounds. These hypotheses are borne out by their empirical

tests (see Section 5). Their study cautions against looking at policy choices in isolation since one

barrier may affect other modes of attack, not initially intended to be affected.

Terrorist groups are more likely to locate in states that are politically unstable or are in

closer proximity to the venues of (planned) attacks (Bapat 2007 and Gaibulloev 2015).19 Failed

and weak states enlarge the pool of likely recruits, create favorable environment for groups to

operate freely, and limit government – domestic and foreign – responses. Venue distance affects

the cost related to supplying terrorist assets from the groups’ home base. In an empirical study,

Gaibulloev (2015) investigates terrorist groups’ locational choice by applying a conditional logit

model with a rich set of controls and some heed paid to endogeneity concerns. He shows that

terrorist groups were more inclined to locate in countries where incumbent terrorist groups

shared a similar ideology to that of the entrant. This was especially true for left-wing terrorist

19 Bapat (2007) also finds that groups choose to locate in a foreign country that is more capable than the target country and has a lower political affinity to the target country.

20

groups. Apparently, a greater number of groups improve their productivity and resourcefulness

through collaboration. Group clustering also forces the resident government to dilute

counterterrorism assets, which bolsters groups’ productivity.

Another terrorist choice involves the venue country. Large-scale efforts to secure borders

following 9/11appears to have transferred attacks away from industrial countries’ soil in North

America and Europe (see Figures 7-10; Enders and Sandler 2006). The transference of attacks

may also concern a terrorist group’s choice of target type as select targets within countries are

fortified. Figure 11 displays the cumulative number of transnational terrorist attacks by the four

target types (i.e., officials, business, private parties, and military) for 1968–2016.20 In the late

1960s and 1970s, terrorists directed most of their attacks against official targets. As these targets

became more protected, terrorist redirected their attacks, in part, to business targets. By the start

of the 1980s, businesses deployed increased protection for their people and property, thereby

leading transnational terrorists to increasingly target private parties as seen in Figure 11. By

1999, cumulative attacks against private civilians surpassed those against officials. Terrorists

display a rational transference of targets with the hardest-to-protect targets becoming the most

vulnerable today (Brandt and Sandler 2012). This is in keeping with our recurring theme that

well-intentioned policies may have undesirable consequences. A similar diagram, not displayed,

applies to domestic attacks with the same sequence of transference, in which target substitution

is much faster as domestic groups learn more quickly from one another (Gaibulloev, Sandler, and

Santifort 2012).

[Figure 11 near here]

3.3 Determinants of Terrorist Groups’ Survival

20 This figure updates one in Brandt and Sandler (2010).

21

Recently, an empirical literature of the duration of terrorist organizations is emerging.21 In an

initial study of groups’ longevity, Jones and Libicki (2008) examine 648 terrorist groups that

were active between 1968 and 2006.22 Blomberg, Engel, and Sawyer (2010) later apply survival

or duration models to ascertain the empirical determinants of terrorist group longevity. Among

their findings, survival in the ensuing period increases as the group ages and its lethality

increases, especially after 1990. This pioneering study suffers from numerous concerns: the sole

reliance on transnational terrorist attacks, the use of targeted countries’ socio-economic

variables, no good measure of the groups’ endpoint, and the absence of tactical controls (e.g.,

groups’ diversity of attacks and their share of transnational attacks).

These particular shortfalls are addressed by Blomberg, Gaibulloev, and Sandler (2011)

and Gaibulloev and Sandler (2013). First these studies provide a theoretic foundation in terms of

a terrorist-choice model among tactics and other considerations. Second, they include both

domestic and transnational terrorist attacks because most terrorist groups engage in both kinds of

attacks. Third, these studies use terrorists’ base-country, socio-economic variables since groups’

survival more likely hinges on conditions where they train and seek refuge. Fourth, measures of

groups’ tactics are included. Fifth, the group data are carefully cleaned to eliminate duplicates

and other concerns (e.g., generic designations). Using continuous- and discrete-time survival

regressions, these studies confirm Jones and Libicki’s findings that religious ideology and group

size are positively associated with terrorist groups’ duration. Greater size improves longevity by

helping terrorist groups stage more successful attacks and endure the loss of leadership.

Terrorist organizations choose large-scale attacks only when they grow sufficiently in size

21 Systematic datasets on terrorist groups were not available until recently. Therefore, earlier studies on survivability of terrorist organizations are mostly qualitative. For example, Cronin (2006) examines survivability of al-Qaida based on historical lessons on evolution and demise of modern terrorist groups. 22 The dataset includes groups that started before 1968 but did not end before 1968. Jones and Libicki (2008) identify the end year of a terrorist group as the year when the group either stops using terrorism or ceases to exist.

22

(Feinstein and Kaplan 2010). Findings also suggest that groups that rely more on transnational

terrorist attacks are more likely to fail, because such attacks put a terrorist group in greater peril

when borders have to be crossed. Furthermore, targeted foreign countries may either retaliate

against terrorists or provide counterterrorism assistance to governments combating the terrorist

group. Finally, a group’s cooperative ties with other terrorist organizations are found to improve

the group’s survival prospects, especially in strong or less democratic states (Phillips 2014).23

Terrorist groups end for a variety of reasons. Some groups exit through political

negotiations and integrate into society, whereas others are defeated through military force.

Understanding what factors lead groups to end in alternative ways is important for effective

counterterrorism policy. Gaibulloev and Sandler (2014) use a competing-risk approach to

distinguish the determinants of three types of endings – internal splintering, forceful elimination,

and joining political process or achieving victory (e.g., achieving some of the goals). Three

important findings emerge. First, religiously motivated terrorist groups are less likely to join the

political process or end victoriously, compared to nationalist/separatist, left-wing, and right-wing

groups. Second, terrorist organizations seeking grandiose goals – i.e., empire, regime change, or

social revolution – and pursuing territorial change are also less likely to end through political

process or victory. Such goals leave very little room for compromise between parties. Third,

having multiple bases of operation reduces the probability of ending by military force or

splintering; eliminating such groups requires collective action by multiple countries, which

seldom occurs. Furthermore, multiple home bases allow for accommodating various viewpoints,

which deters splintering and increases resilience. Some Islamic terrorist groups, such as ISIS

and al-Qaida, possess multiple bases. When Gaibulloev and Sandler (2014) partition the data for

groups starting before and after 1990, the data are too thin for a competing-risk analysis.

23 See, e.g., Gaibulloev and Sandler (2013) and Phillips (2014) for other potential determinants of groups’ longevity.

23

Safe haven and groups’ age are also important. Surprisingly, when a terrorist group is

provided safe haven by a foreign supporter, the likelihood of military defeat increases (Carter

2012). This follows because a host government must balance between its support to terrorists

and the risk of military confrontation with a targeted country. Terrorist groups are at high risk

for ending internally in their infancy; the probability of internal dissolution, however, reduces

significantly as a group lasts for ten years (Carter 2012). Repression may improve groups’

survival prospect in democracies by weakening counterterrorism cooperation from local

communities, expanding popular support and potential recruits, and diminishing a government’s

overall effectiveness in fighting terrorism (Daxecker and Hess 2013). In nondemocracies,

however, repression may worsen groups’ duration insofar as the governments are less

constrained in their efforts to combat terrorism, exercise more control over the media, and are

less concerned about perceived legitimacy of their counterterrorism measures. Daxecker and

Hess (2013) empirically confirm that repression has a negative effect on groups’ demise in

democracies and a positive influence on groups’ failure in nondemocracies. Additionally,

government repression improves the chance of victory for groups in democracies and reduces the

chance of victory in nondemocracies.

Several empirical challenges require attention for the further study of group demise.

First, the quality of data on terrorist groups is a concern. The clandestine nature of terrorism

makes it difficult to collect reliable data on groups, thus raising measurement errors and omitted

variables concerns. Information on groups’ characteristics over time, which are important for

understanding groups’ dynamics, is lacking; e.g., Jones and Libicki’s dataset ends in 2006 and all

group-specific information are time-invariant. Second, additional studies are needed to tackle

causality issues. For example, groups’ size is found to be one of the most robust correlate of

survivability. However, the information on group’s size is based on estimates of a group’s peak

24

size and does not vary over time. The time-invariance of group size does not allow controlling

for fixed effects. In addition, the effect of size on longevity may capture the influences of other

group-specific variables – e.g., group’s financial resources – not available to researchers. These

concerns raise potential endogeneity issues. Third, an expanded dataset on groups, which is

carefully cleaned and merged with an up-to-date terrorist event dataset, is needed. The creation

of the requisite global group dataset that goes to the present day and can be merged with GTD is

a large undertaking. Such data would allow researchers to contrast the changing nature of

terrorist groups before and after 9/11and focus better on Islamic terrorist groups.

3.4 Terrorist Recruitment and Group Organization

During the reign of leftist terrorists in the 1960s, 1970s, and 1980s, groups attracted members

with strongly held political views. These members were prepared to engage in violent acts at

great peril to them for the cause (Crenshaw 1981 and Hoffman 2006). This political-based

membership motivation is still true for left-wing, right-wing, and nationalist/separatist terrorist

groups, but is less descriptive of religious fundamentalist terrorists. Based on interviews and

terrorist profiles, Sageman (2004, 2008) concludes that the latter terrorist groups rely on kinship,

long-term friendships, and worship for recruiting purposes. Such ties are very tight and make it

extremely difficult for the authorities to infiltrate these groups. Additionally, these ties provide a

sense of camaraderie among members that facilitates volunteers for dangerous and even deadly

operations (Wintrobe 2006).

By drawing on the theory of clubs (Cornes and Sandler 1996), the sociology of religion,

and the political science of insurgencies, Berman and Laitin (2008) provide an insightful analysis

on religious terrorist groups’ recruitment, effectiveness, and organization. Their analysis

indicates how Islamic extremist groups limit defections, execute suicide missions, and recruit

25

committed members. Such groups include Hamas, Hezbollah, Muqtada al Sadr’s Mahdi Army,

the Taliban, the Egyptian Muslim Brotherhood, and the Palestinian Muslim Brotherhood. At the

foundation of their theory is the groups’ provision of excludable social services or club goods to

their members and members’ families.

This club model of religious fundamentalist groups is inspired by the work of Iannaccone

(1992), who argue that Christian denominations attempted to limit choices and resources among

their members to obtain their loyalty in exchange for excludable social services. The acceptance

of the religious sects’ mandated prohibitions and sacrifices translates into a “costly signal of

commitment” by their members (Berman and Laitin 2008, p. 1943). By organizational design,

sect members have limited outside opportunities and are isolated from the outside world. With

more stringent prohibitions and sacrifices, these religious groups must offer even larger levels of

club goods to their ever-more dependent members. These club goods serve an essential

recruitment and retention purpose. Berman (2000) extends this club theory to ultra-orthodox

Jews, while Berman (2009) tailors it to radical Islamic terrorist groups. In keeping with

Iannaccone, Berman (2009) also indicates how religious fundamentalist groups induce such

sacrifices from members by sending “their children to religious schools that provide poor

training for employment after graduation” (p. 22).24 In so doing, group members have little or no

economic opportunities outside the club when asked to make the ultimate sacrifice. Members’

religious training also limits outside contact, thus curbing potential leaks prior to an operation.

Berman and Laitin (2008) go on to argue that group solidarity and altruism may motivate

these club members to commit suicide terrorist attacks against hardened target, for which suicide

operations may be the sole means to wreak havoc. In his interview-based analysis of failed

24 This view is more fully developed by the seminal works of Berman (2000), Iannaccone (1992), and Iannaccone and Berman (2006).

26

suicide terrorists, Merari (2004) also characterizes terrorists as altruists, who are not driven by

heavenly rewards, depression, or revenge, contrary to some conventional wisdom.25 The better

equipped is the terrorist group’s adversary, the greater is the group’s need for suicide missions to

penetrate defensive perimeters. A suicide operation provides the added benefit that the

perpetrator of successful missions cannot later offer intelligence on the group. Mission success

is fostered through the commitment of its member operatives, who benefit from group-provided

club goods. As Berman and Laitin point out, suicide missions were deployed against hard

targets during the Second Intifada in Israel, the Iraqi insurgency against Coalition forces, the

Chechen rebellion against Russia, and other hard-target scenarios. Often, but not always, the

targets of these attacks were hardened military or official targets rather than softer civilian ones.

Members’ or operatives’ commitments are stronger for devout individuals with low wages. To

achieve this type of membership, these religious terrorist organizations exclude less religious,

high-wage individuals from their club-provided goods (Berman and Laitin 2008, p. 1952).

These authors maintain that service provision and poor market opportunities allow these

Islamic terrorist groups to motivate their members for effective missions. This leads to two

policy recommendations. First, governments should provide these social services to break these

groups’ monopoly provision. Second, better labor market opportunities are needed to discourage

youths from joining these organizations. In Section 5.4, we briefly discuss empirical support that

these authors provide for some of their hypotheses – e.g., service-providing terrorist groups are

associated with more lethal attacks and suicide attacks are directed at hard targets.

In other situations, terrorist recruitment can be motivated by backlash to harsh

25 Azam (2005) also views suicide terrorists as altruists, making the ultimate sacrifice for the good of future generations.

27

governmental proactive measures leveled against a terrorist group.26 In fact, terrorists seek to

push governments to overreact to a brutal terrorist attack in order to recruit more members

through terrorist supporters’ backlash (Enders and Sandler 2012 and Hoffman 2006). This tactic

applies to all types of terrorist groups. Recruitment increases following successful terrorist

attacks27 or failed, highly visible attacks (e.g., the 1972 Munich Olympics kidnapping and

murder of Israeli athletes by Black September).

A somewhat related recruitment method concerns an “outbidding” strategy among rival

terrorist groups for attracting operatives and constituency support. Bloom (2005, pp. 94-7) view

the ratcheting up of violence among competitive groups to be greater when there are more rival

groups. Suicide terrorist attacks may, in particular, be used for outbidding purposes. To support

her hypothesis, Bloom presents some data plots of terrorist attacks for Hamas, PFLP, Fatah, and

Palestinian Islamic Jihad (PIJ) during 2000–2002 for the Palestinian-Israeli conflict. She also

discusses the rise of the Tamil Tigers over its rival groups in Sri Lanka once the Tigers resorted

to suicide attacks. Bloom’s outbidding hypothesis generated some subsequent empirical tests.

For example, Nemeth (2014) find weak evidence in support of this hypothesis when he examines

domestic terrorist incidents during 1970–1997. Group competition is captured by the inverse of

the Herfindahl-Hirschman Index of market shares. In Nemeth (2014), group competition is

negative and significant, which is consistent with terrorist group free riding rather than

outbidding. When he examines a narrower sample of groups with similar ideologies, and focuses

on the interaction term capturing intergroup competition and society’s acceptance of violence,

there is weak evidence of outbidding. Subsequently, Conrad and Greene (2015) investigate the

severity of attacks rather than the quantity of attacks. Severity increases as terrorists target

26 On backlash, see Arce and Sandler (2010), Dugan and Chenoweth (2012), Rosendorff and Sandler (2004, 2010), and Siqueira and Sandler (2007). 27 See Faria (2003) and Faria and Arce (2005) for a purely theoretical analysis of recruitment in a dynamic model.

28

civilians rather than infrastructure or official targets. Their outbidding evidence is weak with

significance at only the 0.10 level. To date, the most convincing evidence of outbidding is

associated with the micro-level study of the Second Intifada and competition within religious

groups (Hamas and PIJ) and within secular groups (Fatah and PFLP) (Jaeger et al. 2015). These

authors present a multinomial logit choice model where Palestinians indicate their support

among rival factions. There is some evidence that violence increases support within Islamist or

within the secular factions, so that outbidding is ideology specific. Unlike earlier outbidding

articles, Jaeger et al. (2015) provide an identification strategy using district dummies, time

dummies, and district time-varying characteristics.

Since the rise of Islamic extremist terrorism in the 1990s, there have been noteworthy

instances (London, Manchester, Paris, and San Bernardino) of first and second generation

Muslim immigrants attacking domestic targets after undergoing radicalization. Krueger (2008)

examines 63 alleged homegrown Islamic terrorists in the United States. He finds that these

terrorists were younger and better educated than a representative sample of over 1000 Muslim

Americans. Notably, Krueger (2008, p. 295) does not uncover persuasive evidence that sample

homegrown terrorists had been recent converts to Islam. However, his analysis does not

consider the factors behind radicalization. Hegghammer (2013) builds an original dataset of

jihadists in North America, Western Europe, and Australia for 1990–2010. His sample of 401

terrorism plotters includes domestic and foreign fighters, the latter of which had acquired

experience in a foreign theater. The largest proportion of sample Western jihadists favored

fighting abroad over engaging in domestic terrorist attacks. Hegghammer (2013, p. 6)

hypothesizes that this choice reflected jihadists’ preference, smaller barriers, and training

opportunities associated with fighting abroad. When jihadists return from a foreign theater, they

are more adept at domestic terrorist attacks. Hegghammer only displays simple percentages and

29

provides no statistical inference. The determinants of radicalization and, thus, recruitment are

not analyzed. To address the radicalization question, more data are required that include

demographic and socio-economic characteristics of the jihadists.

3.5 Control within Terrorist Group

With few exceptions,28 the literature treats terrorist groups as a unitary entity, which, among

other things, facilitates game-theoretic and empirical analysis. Arce and Siqueira (2014),

Shapiro and Siegel (2007), and Shapiro (2007, 2013) examine the control problem within

terrorist groups wherein a leader (the principal) must delegate attack planning, funds solicitation,

and recruitment to middlemen (agents), who may hold different views than leaders with respect

to the organization’s goals. According to Shapiro (2013, p. 4): “The terrorist’s dilemma is

simple: leaders need to control how violence is executed and how finances are managed, but the

tools to do so create some measure of operational vulnerabilities and therefore increase the

likelihood of operatives being caught and a group compromised.”

Control problems involve at least two tradeoffs: (i) operational security and financial

efficiency and (ii) operational security and logistical efficiency (Shapiro 2013, p. 8). To achieve

greater resource efficiency, the groups must institute greater audits, but this compromises the

group’s secrecy. Agency problems also plague tactical aspects as agents may work at cross-

purpose with leaders. Shapiro (2007, 2013) highlights adverse-selection and moral-hazard

issues. In terms of adverse selection, terrorist operatives, who survive long enough to become

middlemen, may be less committed and less inclined to assume risks. Moral hazard arises

because agents can profit from information asymmetries at the expense of the organization.

28 A noteworthy exception is Siqueira (2005), who investigates conflict and cooperation between the military and political wings of terrorist groups. Another important exception is Bueno de Mesquita (2005a) who distinguishes moderate and hardline factions and their implications for attacks following government concessions.

30

Given security concerns, an optimal incentive-compatible contract would be ill-advised. Such

organizational concerns mean that terrorist groups are less formidable than in the absence of such

incentive and informational considerations. This, in turn, limits somewhat the required level of

counterterrorism.

A final organizational issue involves terrorist network structure, which creates a tradeoff

between vulnerability and operational capability (Enders and Jindapon 2010 and Enders and Su

2007). After 9/11, terrorist groups relied more heavily on loosely linked networks, where a

minimal number of operatives are linked, so that group infiltration compromises only a few

operatives (Memon et al. 2009). This loosely linked structure limits the group’s ability to

perform complex, large-scale missions, which is good for the targeted public. Tighter linked

groups tend to reside in failed states, where more complex missions occur since 9/11 (George

2018).

To date, the control problem within terrorist groups is not subject to empirical tests owing

to data issues.

3.6 When Are Terrorist Groups Effective?

To judge terrorist groups’ effectiveness, one must devise a definition of this effectiveness.

Abrahms (2006, p. 46) specifies two potential effectiveness measures. Combat effectiveness

refers to the group’s induced level of damages, which increases the group’s visibility with its

audience. Such presence can raise the public’s awareness of the group’s political goals or

grievances, and can induce the public to pressure the government to grant concessions. By

contrast, strategic effectiveness refers to the group’s ability to partially or wholly obtain its

policy objective(s) – e.g., Hamas’ or the PIJ’s objective of an autonomous Palestinian state.

Combat effectiveness may ultimately affect strategic effectiveness, but the few effectiveness

31

studies concentrate on the latter.

To date, the evidence on strategic effectiveness is mixed. Abrahms (2006) examines 28

terrorist groups on the US list of foreign terrorist organizations and finds that these groups

obtained their policy objectives only 7% of the time. In his study, limited success – e.g., al-

Qaida’s demand for the removal of US troops from Saudi Arabia – is not counted as a success.29

Because sample terrorist groups seldom achieved their stated policy goal in his study, Abrahms

(2008) subsequently questions these groups’ rationality. Clearly, Abrahms (2006, 2008) raises

an issue of considerable importance; however, his analysis is not complete and offers no

statistical inference. His sample is small and rather selective; it includes only contemporary

groups that the United States designated as terrorist groups. Additionally, Abrahms’ measure of

success is very stringent.

In earlier studies of suicide terrorist campaigns, Pape (2003, 2005) shows that six of

eleven sample campaigns achieved some significant policy concessions, usually resulting in the

withdrawal of foreign forces. The successful cases included concessions granted by democracies

– e.g., the United States, France, Turkey, and India – to terrorist groups. As Abrahms (2006)

points out, the findings from suicide campaigns cannot necessarily be generalized to all terrorist

campaigns. By focusing on one type of campaign, Pape’s (2003) study has a worrisome

selection bias (Asworth et al. 2008).

Using many of Jones and Libicki’s sample groups, Gaibulloev and Sandler (2014, p. 33)

indicate that 23% joined the political process or achieved victory – see Section 3.3.30 In the

former category, these groups obtained a legal avenue to pursue their political agenda and, in the

latter category, they achieved their goal. This success rate may be small, but it is not nil.

29 Al-Qaida demanded the removal of US troops from the Persian Gulf area, including Saudi Arabia. 30 In Jones and Libicki (2008), 132 of 648 terrorist groups are in this combined category.

32

Currently, Gould and Klor (2010) offer the most comprehensive study of terrorism

effectiveness with respect to achieving a group’s political goals. For Israel during 1988–2006,

these authors utilize geographic variation in terror attacks and local voters’ attitudes to show that

Israeli voters in attack-prone areas were more apt to favor granting territorial concessions to the

Palestinians. The authors estimate a nonlinear relationship between the level of local attacks and

Israeli support for such concessions. Up to an attack threshold, an increase in localized terrorism

augmented voters’ willingness to grant concessions; past the attack threshold, further increase in

terrorism inhibited voters’ willingness to grant concessions. These authors also find that as

terrorism rose, voters increased their support of right-wing parties that were gravitating leftward

with attacks.31 Gould and Klor (2010) exercise care with their identification strategy. By

examining only attacks in Israel, these authors do not have to control for numerous country-

specific factors among heterogeneous countries that make causal inference difficult for large n-

country studies concerning terrorism effectiveness. In addition, these authors show that reverse

causality, wherein voters’ attitudes influence local terrorist attacks, does not hold. Their reliance

on subnational micro-level data over 19 years gives a great deal of variation.32 Unlike some

terrorist campaigns, the goals of the Palestinian terrorists are very clear, which also fosters

inference.

Although the question of terrorist groups’ effectiveness is by no means settled, there is

mounting evidence that these groups, at times, do achieve some of their goals, particularly when

these goals are not too grandiose.

4. Game Theory and Empirical Studies of Terrorism 31 Also see Berrebi and Klor (2006) and Gassebner, Jong-A-Pin, and Mierau (2011) on terrorism and elections. 32 Jaeger et al. (2012) also use micro-level data of opinion polls to investigate how Israeli violence against Palestinians bolstered more political support for Palestinian terrorist groups. This support provided the necessary resources for a sustained terror campaign that pressured Israel to make concessions.

33

Game theory or strategic rational choice has been applied to the study of terrorism. Early

contributions include Selten’s (1977) study of kidnapping and Sandler, Tschirhart, and Cauley’s

(1983) study of terrorist-government negotiations. Noncooperative game theory is particularly

ideal for analyzing counterterrorism because targeted governments and terrorists must employ

actions to foster their self-interest subject to constraints, whose parameters are, in part,

determined by the response of their counterparts. For example, a targeted government’s actions

against a common threat from a terrorist group hinge on the actions being taken by other targeted

governments. Additionally, a terrorist group’s resource constraint depends on the government’s

proactive measures leveled at the group’s personnel and assets. Our intent is to indicate in select

instances how game theory informs policy and empirical studies of terrorism.33

4.1 Making Concessions to Terrorist Hostage Takers

Grim images of ISIS hostages in orange jump suits stoically waiting to be beheaded after their

American and British governments refused to bargain for their release raises a heart-rending

policy issue about paying ransoms. This issue is even more poignant when hostages from Spain,

France, Belgium, Italy, and elsewhere have been released after ransom payments to ISIS (Brandt,

George, and Sandler 2016).34

The proper reaction to hostage-taking terrorists raises a strategic move on the part of a

government. If terrorists are solely motivated to abduct hostages to gain funding or the release

of their comrades in order to support their terrorist campaign, then a credible threat by a

government never to concede to terrorist hostage takers may dissuade terrorists from abducting

33 Sandler and Arce (2003) and Sandler and Siqueira (2009) offer surveys on game theory and terrorism. Also see Schneider, Brück, and Meierrieks (2015). 34 On empirical studies on hostage negotiations, also see Atkinson, Sandler, and Tschirhart (1987), Gaibulloev and Sandler (2009a), and Santifort and Sandler (2013). Bueno de Mesquita (2005a) investigates negotiations between terrorists and a government over general demands, not involving hostages.

34

that government’s citizens. Lapan and Sandler (1988) investigate the time inconsistency of such

a pledge in which the abduction of a sufficiently valuable hostage (e.g., a soldier, diplomat, or

government official) may cause the government to renege on its pledge. The no-concession

pledge may fail to deter hostage taking if terrorists’ net gains from logistical failure (i.e., not

securing intended hostages) or negotiation failure (i.e., failing to receive a concession) are

positive. Martyrdom may yield a positive gain for a logistical failure, while kidnap-generated

publicity or induced societal fear may offer a positive payoff for negotiation failure. ISIS

Internet-posted beheadings have produced such fear and have resulted in the recruitment of

fighters in Iraq and Syria. A government’s no-concession pledge only ends all such hostage

taking if all of the following holds: the pledge is credible and irreversible, the terrorists only

benefit from concessions, and the no-concession consequences for the government is known

(Lapan and Sandler 1988). Because the identity and, thus, value of the hostage cannot be known

beforehand, the consequences from not conceding are necessarily uncertain. Thus, the efficacy

of this strategic move is open to empirical observation.

Based on transnational terrorist hostage incidents during 1968–2005, Brandt and Sandler

(2009) apply refined time series estimating procedures to test the conventional wisdom about the

efficacy of a no-concession policy. Based on monthly and quarterly hostage data, their event

count method reduces bias and augments consistency when estimating cyclical and structural

components of the hostage-taking series, compared to standard forecasting or linear regression

methods. To establish the necessary baseline, their analysis first identifies changepoints to the

hostage-taking data that arise from exogenous factors (e.g., the deployment of metal detectors to

airports in 1973 and the end of the Cold War) for each of the hostage-taking series (i.e.,

skyjackings, kidnappings, and other hostage events).

Brandt and Sandler (2009) identify the lag structure of the three (monthly and quarterly)

35

hostage-taking series. In so doing, they estimate the short- and long-run impact multipliers

associated with an additional negotiation success for the terrorists. For kidnappings, the long-run

multiplier is 2.62 additional incidents; for skyjackings, the long-run multiplier is 0.59 additional

incidents. The long-run multiplier for other hostage incidents is insignificant (Brandt and

Sandler 2009, p. 770). Hence, for kidnappings and skyjackings, the conventional wisdom is

deemed valid: giving into terrorist demands results in more such incidents. Moreover, these

concessions can fund other operations of the terrorists.

For the period studied by Brandt and Sandler (2009), there were fewer religious

fundamentalist terrorist groups than after 2005. Also, the United States was much less consistent

in maintaining its no-concession pledge prior to 2001 than after 2001. During 1978–2000, the

United States gave concessions to kidnappers 23.1% of the time, whereas during 2001–2013, it

gave concessions to kidnappers 10.7% of the time (Brandt, George, and Sandler 2016, p. 46).

ISIS and al-Qaida in Iraq, which used beheadings, surfaced during 2001–2013. Brandt, George,

and Sandler (2016) apply their earlier methodology to three samples: (i) the US-UK that

generally stood by their no-concession pledge; (ii) ten concessionaire countries that less

frequently stood by their no-concession pledge; and (iii) the EU countries, less the UK, that less

frequently stood by their no-concession pledge. These authors find that making a concession

increased the median baseline kidnappings by 64 to 87%, depending on the cohort of countries

and the particular changepoint regime. Once again, the message is unmistakable: making

concessions increased the level of kidnappings. For this second study, these authors only

examined kidnappings, which became the lion’s share of hostage incidents after 9/11.

4.2 On US Military Aid, Negotiations, and Moral Hazard

Our second example not only indicates how game theory can inform counterterrorism policy, but

36

also shows how strategic factors can result in perverse incentives as in the previous case. In

particular, we examine Bapat’s (2011) provocative analysis of US military aid to countries

hosting terrorist groups. In his underlying game-theoretic model, there are three players: the

United States, the recipient (host) country, and the host’s resident terrorist group. The United

States first chooses its level of aid. The host country then decides whether to negotiate with the

terrorists, to defend against them, or to mount an offensive to defeat them. If there is a

negotiation, the terrorists must accept or reject the host country’s offer. When the offer is

accepted, the United States must decide whether or not to continue its aid. The game can

conclude in three ways: the host government loses in a battle with the terrorists; the host

government and the terrorists reach an accommodation; or the terrorists are defeated in an

offensive or a failed attempt to seize power. Bapat (2011) solves for the solution to this

infinitely repeated game with incomplete information.

This solution displays the moral hazard problem associated with US military aid to a

government, hosting a terrorist group. Ideally, this aid is intended to assist the host country to

defeat the terrorists, thereby stabilizing its regime while limiting the risks of terrorist attacks on

US interests. However, the aid sets up a perverse incentive to the host country not to defeat the

resident terrorist group in order to keep the aid coming. As Bapat (2011) points out, large

military aid flows to the Musharraf regime in Pakistan after 9/11 resulted in little Pakistani action

against the Taliban. Similar scenarios were true in post-9/11 Afghanistan and Yemen where

little were done by recipient governments to defeat their resident terrorists. Despite this moral

hazard concern, Bapat’s (2011) solution shows that the United States still profits from the aid

because it inhibits an accommodation between the host and the terrorists, thereby maintaining the

US-favored regime status quo.

Bapat (2011) provides an empirical test of his model based on a hazard model,

37

corresponding to the duration until the resident terrorists are defeated. His dataset contains 184

terrorist campaigns against 48 host countries during 1997–2006. He finds that US military

assistance increases the duration of a resident terrorist group in an aid-recipient host country

from a mean of 4.69 to 7.82 years, a 67% increase! This is another example of an unintended

negative consequence from well-intended counterterrorism policy. He also shows that host

countries with a larger affinity to the United States (i.e., closer voting record on the UN General

Assembly with US-held positions) display further increases in the duration of their resident

terrorist groups to 13.38 years. This follows because the United States is more motivated to keep

its military aid flowing to these supportive countries even if the terrorist threat remains. Cutting

off aid may result in a host-terrorist accommodation that ends support of US policy positions.

Bapat’s (2011) theory-based empirical test explains the often-observed failure of military aid to

eliminate resident terrorist groups.

4.3 Game Theory and Counterterrorism Effectiveness

Game theory informs the empirical analysis of counterterrorism practice by indicating the likely

choice of defensive and proactive measures by commonly targeted governments.35 In other

instances, the game involves the endogenous interaction among targeted government(s) and the

terrorist organization (Cárceles-Poveda and Tauman 2011, Jindapon and Neilson 2009, Powell

2007, Siqueira and Sandler 2006, and Zhuang and Bier 2007). Moreover, the players may be

uninformed about their adversaries’ type or resources (Arce and Sandler 2007, Bier, Oliveros,

and Samuelson 2007, Lapan and Sandler 1993, and Overgaard 1994). And in still other

instances, players include a targeted government, a terrorist organization, potential terrorist

35 Key theoretical articles include Arce and Sandler (2005), Kunreuther and Heal (2003) Sandler and Lapan (1988), and Sandler and Siqueira (2006). The latter forms the basis for the described model in the text. Propositions from some of this literature are tested in the laboratory – see Arce, Croson, and Eckel (2011).

38

recruits, and/or voters (Bueno de Mesquita 2005b, 2007 and Siqueira and Sandler 2007, 2010).

Our intention is not to review the game-theoretic literature on counterterrorism, but rather

to discuss briefly strategic aspects of defensive and proactive measures by commonly targeted

countries. This model explains overspending on such measures as governments seek to transfer

transnational terrorist attacks to foreign venues. Among targeted countries, this regional

transference of terrorist attacks shows up before and after 9/11 in Figures 7 and 9 in terms of

incidents and casualties, respectively. Enders and Sandler (2006, 2012) show the geographical

transference of transnational terrorist attacks from the Western Hemisphere and Europe to

Africa, Asia, and the Middle East. Because there is no requisite panel defense data, researcher

must infer this transference after a watershed event, such as 9/11 that is believed associated with

greater defensive spending in the West. Among target types, an analogous transference is

evident in Figure 11, as official and business targets are hardened.

Consider a scenario in which two countries are targeted by the same terrorist group that

can strike either country’s interests at home or abroad. It is reasonable to assume that the same

magnitude attack is more devastating to a country’s interests at home than abroad. At home,

defensive measures against transnational terrorism attacks not only limit the consequences of

such an attack, but also decrease its likelihood. The latter arises as the terrorist group seeks a

softer target abroad. Greater defense in either country may deflect an attack abroad where both

countries’ assets are harmed to varying degrees.36 Such defensive countermeasures imply at

least two opposing externalities. The more important one is the negative consequences of

transferring a terrorist planned attack abroad to a less protected country (Sandler and Lapan 1988

and Sandler and Siqueira 2006). A smaller positive externality from home defense is the

36 Conceptually, the terrorist group is a passive agent that seeks out the most vulnerable country with the highest anticipated payoff. Cárceles-Poveda and Tauman (2011) allow the terrorist group to be an active agent. The transference result still holds with an active terrorist group as targeted countries work at cross-purposes.

39

protection afforded to foreign-held assets.

Given the dominant transference externality, each of the two targeted countries’ Nash

reaction paths, relating the country’s own defense choice to that of the other country, is

positively sloped, consistent with strategic complements and a defensive race. The latter leads to

overprovision of defense compared to the social optimum level.37 To some extent, the protecting

country’s foreign interests curb the motive to transfer attacks abroad through a fortification race

(Bandyopadhyay and Sandler 2011, 2014), but this curb is attenuated given that a country’s

home interests are more important than its foreign interests. The presence of transference

externalities indicates the need for international cooperation regarding the choice of defensive

countermeasures against a common terrorist threat.

Next, we consider proactive measures (e.g., retaliatory raids, drone attacks, or group

infiltration) by either targeted country against the common terrorist threat. Such actions decrease

the likelihood and consequences of a transnational terrorist attack at home and abroad. As such,

proactive measures against this common threat give rise to positive externalities to the other

targeted country, given the purely public nature of these actions. Now, the associated Nash

reaction paths are negatively sloped, consistent with strategic substitutes for which too few

measures are provided by the targeted countries.

If many countries are in the terrorist group’s cross hairs, then the country sustaining the

greatest losses is the likely one to confront proactively the terrorists, as the United States did

against the Taliban and al-Qaida following 9/11 (Arce and Sandler 2005). Nevertheless, the

proactive country is not anticipated to internalize fully the positive externality that its offensive

confers on other at-risk countries. If we evaluate the cooperative solution at the Nash

equilibrium level of proactive measures, underprovision is implied. Once again, there is a clear

37 This analysis can be readily extended to more targeted countries (Cárceles-Poveda and Tauman 2011).

40

role for international cooperation.

Other factors can enrich the game-theoretic analysis of counterterrorism. First, we can

allow for leader-follower behavior, which worsens proactive underprovision and but limits

defensive overprovision (Sandler and Siqueira 2006). Second, for proactive measures, backlash

can be included in which these actions create new grievances that can offset benefits derived

from weakening the terrorist group (Das 2008, Rosendorff and Sandler 2004, 2010, and Siqueira

and Sandler 2007). Third, both defensive and proactive measures can be chosen together

(Bandyopadhyay and Sandler 2011). Fourth, more agents (e.g., voters, terrorists, and supporters)

can be included in the analysis. Fifth, we can account for more political considerations when

investigating counterterrorism policy (Bueno de Mesquita 2007). Sixth, these game-theoretic

choices can be viewed solely within a country besieged with domestic terrorism and needing to

protect alternative targets or venues, leading to transference among target types as empirically

documented by Brandt and Sandler (2010, 2012). A central government can internalize these

externalities.

5. The Effectiveness of Counterterrorism Policies

To fight terrorism effectively, the world must learn from experience what countermeasures work

and what ones do not. Even before 9/11, there were some noteworthy investigations on this

issue. In a seminal study, Landes (1978) applies a Becker-inspired crime deterrence model to

investigate what best deterred skyjackings in the United States. His study was motivated, in part,

by the observed sharp decline in US skyjackings following the installation of metal detectors in

US airports on January 5, 1973. Using Federal Aviation Administration (FAA) skyjacking data

for 1961–1976, Landes (1978) first presents a linear regression on the number of US hijackings

per quarter based on some potential deterrent measures and opportunity cost variables (e.g.,

41

unemployment rate and GDP per capita). Significant deterrents included the probability of

apprehension and the average sentence of convicted hijackers. Perhaps surprising, neither the

presence of sky marshals nor the death of the hijacker deterred skyjackings.38 In addition,

Landes examines what considerations increased the time interval between hijackings and shows

that this interval increased with the probability of apprehension, the conditional likelihood of

incarceration, and the length of the sentence. Greater income per capita and reduced

unemployment increased the interval between skyjackings (Landes 1978). Based on his

estimated model, Landes (1978, p. 21) computes a counterfactual to show that in the absence of

metal detectors there would have been over 2.7 additional US skyjackings per quarter. Given the

deployment cost of these metal detectors and the estimated benefit from fewer hijackings,

Landes (1978, pp. 21-2) indicates that these metal detectors had a large benefit-cost ratio. This

landmark study indicates that technological-based defensive measures work well against specific

attack modes; however, such barriers can induce terrorists to substitute to other unprotected

attack modes (Enders and Sandler 2004).

Following up on Landes (1978), Enders and Sandler (1993) utilize a vector-

autoregression-intervention (VAR-I) analysis to investigate the substitution-induced

consequences of the installation of metal detectors in airports worldwide during the first quarter

of 1973. Their VAR model allows for interaction among the various hostage-taking and other

attack mode time series, thereby permitting substitution and complement relationships as

terrorists react to these barriers by seeking softer targets. This method allows for the endogenous

choice among modes of attacks and is used by some post-9/11 articles (e.g., Brandt and Sandler

2009). In the short run, these metal detectors reduced skyjackings by just over twelve incidents

38 Sky marshals did not make themselves known for fear that discharging their gun would pierce the fuselage and bring down the plane.

42

per quarter but increased other hostage-taking incidents by about four per quarter (Enders and

Sandler 2012). In the longer run, these other hostage-taking incidents rose by just over five

incidents per quarter. Even more disturbing, an updated study shows that the introduction of

metal detectors increased terrorist incidents with casualties by 8.83 and 12.38 incidents per

quarter in the short and long run, respectively, as terrorists substituted out of skyjackings, which

generally had few casualties before 9/11, into more casualty-rich incidents to capture headlines

(Enders and Sandler 2012, p. 91). Such unintended downsides of metal detectors are not

recognized by Landes (1978).

Enders and Sandler’s (1993) methodology is applied to other defensive measures to

illustrate unintended consequences. For example, the fortification of US embassies resulted in

increased assassinations of diplomatic personnel outside of secured embassy compounds. The

effectiveness of technological barriers fades with time as terrorists devise means to circumvent

them, such as massive truck bombs to get beyond embassy walls or plastic guns and containers

with flammable liquids to get past metal detectors (Enders, Sandler, and Cauley 1990).

Prior to 9/11, the effectiveness of retaliatory raids against alleged state sponsors of

terrorism has been analyzed. Enders and Sandler (1993, 2012) apply their VAR-I method to

ascertain the effectiveness of the US retaliatory bombing raid on Libya on April 15, 1986 for

Gaddafi’s alleged involvement in the La Belle discotheque bombing in West Berlin on April 4,

1986 that killed 3 and wounded 231 including 62 Americans (Mickolus, Sandler, and Murdock

1989, pp. 365-7). Once again, their VAR method permits multiple time series (e.g., attacks

against US and UK interests and attacks against other countries’ interests) to be examined to

account for substitution and complementarity. Also, the timing of reactions is captured. Enders

and Sandler (1993) find that terrorist attacks against US and UK interests skyrocketed following

the raid as terrorists moved planned future attacks into the present to display their outrage. UK

43

interests were targeted since US warplanes had taken off from British airbases. Within a few

quarters, such attacks dropped precipitously as terrorists had to replenish their spent resources.

The overall mean of terrorist incidents did not change owing to the raid. In a similar study of

Israeli retaliatory raids following Palestinian terrorist attacks, Brophy-Baermann and Conybeare

(1994) uncover the same sine-wave pattern, indicative of terrorists’ intertemporal substitution

(Berrebi and Lakdawalla 2007 and Enders and Sandler 2002), in which these raid did not curb

Palestinian terrorism in the medium term.

After 9/11, the sustained War on Terror is seen to have apparently little long-term effect

on global terrorism. This is suggested by the plots of the transnational terrorism incidents and

casualties per incident in Figures 1 and 2 after 2006 and is further supported by the rising

proportion of casualty incidents in Figure 3. A similar message that post-9/11 proactive

measures have a poor payback comes from the counterfactual-based study of Sandler, Arce, and

Enders (2009), where the return to the War on Terror was pennies on the dollar. Furthermore,

enhanced border security since 9/11 caused transference of attacks from North America and

Europe to the Middle East, Africa, and Asia, consistent with the earlier defensive game theory

model (Drakos and Gofas 2006 and Enders and Sandler 2006).

5.1 Drones and Targeted Killings

After 9/11, drones were used by the United States to target terrorist assets and Islamic terrorist

groups’ leadership in Afghanistan, Pakistan, Yemen, and elsewhere. In Pakistan, these drones

were dispatched to seven agencies or districts in the Federally Administered Tribal Areas

(FATA) (Fair 2010). Opponents argue that drones are locally unpopular and create backlash that

fosters terrorist recruitment (Bloom 2005 and Rosendorff and Sandler 2004). Proponents view

effective drone attacks as disrupting and degrading targeted terrorist groups, thus reducing the

44

incidence and lethality of their attacks (Jaeger and Siddique 2011 and Johnston and Sarbahi

2016).

The evidence supports both viewpoints with somewhat more support for limiting terrorist

attacks in the short run following a successful drone strike, especially in Pakistan. For example,

Jaeger and Siddique (2011) apply a multiple time series method, similar to Enders and Sandler

(1993), to show that drone attacks had no significant effect in Afghanistan, but did serve a short-

term deterrent effect in Pakistan. For Pakistan, Jaeger and Siddique (2011) find an increase in

the vengeance or backlash effect during the first week following a successful drone strike on

Taliban and al-Qaida assets. The short-term reduction in terrorist attacks came during the second

week after the drone strike, consistent with an intertemporal transference of attacks. In a

subsequent study, Johnston and Sarbahi (2016) used geocoded data on drone strikes and terrorist

attacks in Pakistan. Their empirical method is a spatial panel, where the unit of analysis is the

district-week. These authors find that terrorist incidents and casualties decreased by 5 and 25

percentage points, respectively, in the week of the drone strike. However, the killing of militant

leaders by drones did not reduce terrorist attacks or their consequences in a significant fashion.

A very interesting additional finding was that there was no apparent transference externality;

namely, the short-term decrease in terrorism did not increase terrorism in neighboring regions.

Moreover, the drone attacks reduced selected assassinations of tribal elders by Islamic

extremists. Long-term permanent disruptive impacts of drone attacks are not addressed. The

dynamics of the effectiveness of drone strikes warrant further investigation.

A related proactive policy is the use of targeted killings of terrorist operatives and leaders

without the use of drones. Once again, such killings may incite more violence through backlash

and recruitment, or alternatively they may limit violence through disruption and deterrence.

Tests of these opposing forces are investigated for the Second Intifada (2000–2005) and beyond.

45

Palestinian violence often assumed the form of suicide bombings of Israeli civilians, while Israeli

violence took the form of targeted killings. These assassinations were often directed at

Palestinian targets where the attacks originated. Jaeger and Paserman (2008) investigate the

causality of the violence, showing that Palestinian suicide attacks incited Israeli retaliation but

that Israeli violence apparently did not induce a Palestinian response in terms of additional Israeli

fatalities. As such, this evidence does not support a tit-for-tat cycle of violence as hypothesized

by Bloom (2005). There is a short-term deterrent outcome in terms of reduced Palestinian

violence coming from Israeli targeted killings.

In a follow-up study, Jaeger and Paserman (2009) examine further the dynamics of

suicide attacks and targeted killings during the Second Intifada. This study treats both sides to

the conflict in a symmetric fashion. Moreover, their data not only account for the location of the

attack, but also the origin location of the perpetrators. This location data better allow for an

assessment of the vengeance versus disruption effect of Israeli targeted killings. These authors

find that a successful suicide attack by Palestinian terrorists resulted in 6.7 more Palestinian

deaths in the following week. Israel reacted to actual, but not planned, suicide attacks. Based on

geocoded data, Jaeger and Paserman (2009) show that targeted killings in the origin district of

the attack reduced subsequent attacks in the short run from that district, consistent with the

disruption hypothesis.

5.2 Thinking Outside the Box

Typically, defensive and proactive measures are put forward as the best ways to counter the

threat of terrorism. Some proactive policies in terms of foiling plots or infiltrating groups can be

quite effective, but there is no global dataset that records such actions so that knowledge of these

efforts is episodic and not subject to rigorous tests.

46

With respect to extremist religious terrorist groups, Berman (2009) puts forward

increases in state-provided social services as an effective means to loosen the grip that these

groups exercise over their operatives and supporters. These radical groups use their “club-

provided” services to elicit sacrifices from adherents (see Section 3.4). In some cases, members

are willing to commit suicide terrorist acts so that their families continue to receive these

excludable services. Berman and Laitin (2008) show empirically that service-providing terrorist

groups (e.g., Hamas and Hezbollah) were associated with more lethal attacks than their non-

service-providing counterparts. Moreover, they find that service-providing terrorist groups

engaged in a higher proportion of suicide attacks. Additionally, these groups used suicide

attacks against hard targets inside the Green Line in Israel. By eliminating the radical group as

the sole provider of these services, the government can break the hold that these groups weigh

over members. This is an interesting proposal that is worthy of empirical test beyond that of

Berman and Laitin (2008) as appropriate data become available.

Dugan and Chenoweth (2012) also think innovatively by investigating whether retaliatory

or conciliatory actions during three violent periods – the First Intifada (1987–1993), the Oslo

Lull (1993–2000), and the Second Intifada – involving Israel and the Palestinians served to ease

or exacerbate terrorism. By raising the expected utility of nonterrorism through rewards,

conciliatory acts may reduce terrorism; 39 by lowering the expected utility of terrorism through

punishment, retaliatory acts may reduce terrorism. Which policy works best to curb terrorism is

an empirical question. During these violent periods, these authors use a text-reading program

applied to Reuter’s archives to code four types of Israeli actions: discriminate retaliation,

indiscriminate retaliation, discriminate conciliation, and indiscriminate conciliation. Based on a

Negative Binomial regression, Dugan and Chenoweth (2012, pp. 614-15) offer evidence that

39 Also, see theoretical pieces by Anderton and Carter (2005), Frey and Luechinger (2003), and Sandler (2010).

47

relatively large numbers of conciliatory acts (i.e., more than eight) and indiscriminate

conciliatory acts reduced terrorism.40 However, they uncover no evidence that indiscriminate

retaliatory acts lowered terrorism. In fact, such repression either had no effect or raised

terrorism, consistent with the notion of backlash. Thus, conciliatory acts – even indiscriminate

types – are more effective than indiscriminate retaliatory acts.

In an innovative exercise, Berman, Shapiro, and Felter (2011) investigate whether service

provision during the Iraq insurgency reduced attacks against occupying forces. For small-scale

reconstruction projects between March 2003 and December 2007, these authors show that

services and counterinsurgency are complements. In fact, the local community helped

counterinsurgency efforts by providing useful strategic information on insurgents’ location. This

information flow increased as more small-scale reconstruction projects were built. Berman,

Shapiro, and Felter (2011) rely on a unique geocoded dataset on violence against US and Iraqi

forces and the location of these reconstruction projects. Once again, the message is that

community projects directed at potential terrorist supporters may curb violence and reduce the

effectiveness of terrorists’ and insurgents’ propaganda (also, see Berman 2009 and Siqueira and

Sandler 2007).

As a deterrent to suicide terrorism during the Second Intifada, Benmelech, Berrebi, and

Klor (2015) examine house demolitions. Punitive demolitions are leveled against the houses of

the family of a suicide terrorist or operatives involved in a suicide attack, while precautionary

demolitions are directed against a region from which a suicide attack may emanate. Unlike

precautionary house demolitions, punitive demolitions are discriminatory. To test the

effectiveness of each type of demolition, the authors join longitudinal micro-level data for 2000–

40 This is in keeping with the notion that kind acts by a terrorism-besieged government may win back the hearts and minds of terrorist supporters – see Jaeger et al. (2012) and Laitin and Shapiro (2008).

48

2005 collected by Israeli authorities on demolitions with data on suicide attacks against Israeli

targets. Their analysis shows that punitive demolitions are effective – a one standard deviation

increase in these demolitions resulted in an 11.7% fall in suicide attacks. In contrast,

precautionary demolitions resulted in backlash for which a one standard deviation increase in

these demolitions ignited a 48.7% increase in suicide attacks! The message is similar to that in

Fearon and Laitin (2003) that countermeasures must be discriminate to be effective (Jacobson

and Kaplan 2007 and Rosendorff and Sandler 2004).41

5.3 Foreign Aid and Counterterrorism

Following 9/11, academic interest grew in ascertaining how foreign aid influenced transnational

terrorism and how the War on Terror affected foreign aid. Because many transnational terrorist

groups, which target Western interests, reside in developing countries, foreign aid can be used by

targeted donor countries to ameliorate its terror threat (Bandyopadhyay, Sandler, and Younas

2011, 2014 and Azam and Delacroix 2006). In evaluating the evidence, one must not lose sight

of the moral hazard problem shown by Bapat (2011) (Section 4.2). Fleck and Kilby (2010)

document changes in US bilateral aid allocations before and after 9/11. According to these

authors (p. 185), US aid rose during the Bush administration; however, need-driven aid fell

relative to other motives. In particular, Fleck and Kilby (2010, pp. 187-8) find that a recipient’s

military budget was a significant determinant of increases in US bilateral aid, suggesting US

interest to support countries that addressed a resident terrorist threat. In an expanded panel study

of 22 donor countries, Dreher and Fuchs (2011) show that official development assistance

(ODA) increased, but not monotonically, during the War on Terror. These authors do not

41 This finding is similar to that of Zussman and Zussman (2006) in which Israeli assassination of Palestinian military personnel resulted in little backlash in contrast to Israeli assassination of senior Palestinian political leaders.

49

discern a significant relationship between aid flows and terrorist incidents. Based on their aid

allocation estimations, Dreher and Fuchs (2011, p. 337) show “that countries where terror

originates are not more likely to receive aid as a consequence, but if they are selected, they

receive larger amounts of aid.” Thus, the War on Terror affected aid flows somewhat, but other

factors – e.g., voting records in the UN General Assembly and sponsoring terror – had a more

profound influence.

The more interesting question is the effectiveness of aid in ameliorating terrorism or its

negative consequences. Most studies focus on aid and transnational terrorism (see, e.g., Azam

and Delacroix 2006 and Azam and Thelen 2008, 2010). This effectiveness question is

intrinsically difficult to identify since terrorism can attract foreign aid and foreign aid may or

may not limit terrorism. Azam and Thelen (2010) exercise care to address this endogeneity and,

by so doing, conclude that foreign aid had a significant effect in curbing transnational terrorist

incidents originating from aid-recipient countries. However, they also find that US military

intervention increases such terrorist incidents. In a follow-up study, Young and Findley (2011)

disaggregate foreign aid by sectors – education, health, conflict prevention, governance, and civil

society – based on AidData. They employ a couple of methods to address endogeneity concerns,

including dynamic GMM models. Young and Findley (2011) find that education, health, civil-

society, and conflict-prevention aid reduced the incidence of transnational terrorism originating

from the aid-recipient country. Conflict-prevention aid had the greatest terrorism-reducing

effect, while governance aid did not have a significant effect. We believe that it is worth

returning to his question to get more robust results as more panel data sectoral aid in terrorism-

plagued countries become available.

5.4 International Cooperation and Counterterrorism

50

For transnational terrorism, the game theory in Section 4 indicates that defensive measures tend

to be overprovided, while proactive measures tend to be underprovided. Thus, there is a need for

international cooperation when addressing common terrorist threats. After 9/11, select countries

joined US-led actions to attack the Taliban and al-Qaida in Afghanistan. More recently, some

NATO allies joined US bombing missions against ISIS targets in Syria – e.g., French bombing

raids on Raqqa, days following the November 13, 2015 ISIS attacks on Paris. On occasion,

targeted countries share intelligence. Nonetheless, most countries address their transnational

terrorist threats independently and, for the most part, international cooperation has not been very

prevalent.

The United Nations and the international community enacted a host of conventions and

resolutions against terrorism and its practices – e.g., taking hostages, skyjackings, and downing

airliners. A recent instance is the International Convention for the Suppression of Terrorist

Bombings, signed in December 1997 and entered into US law on May 2001. Based on time

series intervention analysis, no such UN conventions changed the mean of the outlawed mode of

attack after ratification (Enders and Sandler 2012 and Enders, Sandler, and Cauley 1990), thus

calling into question the effectiveness of such international actions that have no enforcement

mechanism.

Similarly, efforts to cut off terrorist financing through guidance from international

organizations, such as the International Monetary Fund and the Financial Action Task Force on

Money Laundering (FATF), has met with mixed success (Basile 2004). This follows because

terrorists have found ways to circumvent these efforts. Only a few noncompliant banking

institutions can spoil the collective action of compliant banking institutions, since curbing

terrorist financing is a weakest-link public good for all targeted countries (Sandler 2005).

After 9/11, an INTERPOL-surveillance system to secure ports of entry against terrorists

51

has yielded a large apparent payoff at a small cost (Sandler, Arce, and Enders 2011).42 Based on

terrorist arrest data provided by INTERPOL, these authors calculated, under various

counterfactual scenarios, the value saved through INTERPOL-surveillance-assisted arrests.

These savings were in terms of fewer terrorist incidents and the concomitant reduced deaths,

injuries, and property losses. Sandler, Arce, and Enders (2011) build their counterfactuals based

on the casualties and losses associated with an average transnational terrorism incident during

the years prior to the surveillance system being deployed. Value-of-life computations translate

avoided deaths and injuries into a dollar value. In addition, INTERPOL provided these authors

with the cost for the components of the surveillance system. Their exercise shows that every

dollar of INTERPOL’s surveillance expense returned $40 to $200, depending on 12 alternative

counterfactual scenarios. The large return is indicative of the network externality that this

system provides. Thus, countries may reap gains from cooperation that mandates little expense

or loss of autonomy. This example highlights that there are instances of low-hanging fruit with

respect to coordinated efforts to curb transnational terrorism.

6. Three Alleged Causes of Terrorism

After 9/11, the news media and political leaders speculated on the causes of these unprecedented

attacks (Li and Schaub 2004). Three potential causes for transnational terrorism were put

forward: globalization (Sageman 2008 and Schneider, Brück, and Meierrieks 2015, p. 131),

world poverty (Krueger and Malečková 2003), and democracy (Eubank and Weinberg 1994,

2001). We examine each cause in seriatim from the lens of the literature that each generated.

42 See Enders and Sandler (2011) on the factors that influence countries’ decisions to install INTERPOL’s MIND/FIND surveillance system and Gardeazabal and Sandler (2015) on the role of this system in the fight against terrorism.

52

6.1 Globalization and Transnational Terrorism

Economic globalization captures the increased flows of transactions across international borders.

These flows encompass goods, services, financial capital, virtual transmissions, productive

factors, and externalities as countries’ economies become more integrated. Li and Schaub

(2004) indicate that globalization may foster transnational terrorism by curbing the cost of illegal

activities, providing greater cover for transiting terrorists, offering greater financing for terrorism

(e.g., enhanced remittances), supplying transmission channels for terrorist propaganda, and

augmenting grievances through greater income disparity. Globalization may also foster

transnational terrorism through a demonstration effect as increased transactions, real and virtual,

bind countries closer together. Grievance in one country can spill over to another through the

media, the Internet, and exchanges. By contrast, globalization may inhibit transnational

terrorism by improving economic development, thereby reducing potential grievances. In a

dynamic panel study of 112 countries for 1975–1997, Li and Schaub (2004) use trade, FDI, and

portfolio investments as proxies for globalization. Notably, none of the authors’ sample years

are near 9/11. They find no direct linkage between globalization and transnational terrorism.

This initial study on the alleged globalization cause of transnational terrorism never really

addressed the obvious endogeneity concern that transnational terrorism may influence the

authors’ three proxies for globalization (see Section 7). The globalization-terrorism linkage

poses a difficult empirical exercise. Although the Li and Schaub (2004) study is well-cited, it

does not put the question to rest.

In a general study of the causes of terrorism, Gassebner and Luechinger (2011) apply a

modified extreme bound analysis (EBA) to a host of possible determinants of transnational

terrorism including globalization based on Dreher’s (2006) KOF index of globalization. Among

other things, they show that globalization is not a robust determinant of transnational terrorism in

53

the various runs based on venue country, victim country, or perpetrators’ origin. Their

globalization finding agrees with other panel studies.43 Many subsequent articles rely on

Gassebner and Luechinger’s (2011) “robust” variable from their EBA study to come up with a

set of controls when investigating other determinants of terrorism. Thus, it is important that

some essential issues with EBA are raised. Brock, Durlauf, and West (2003) indicate some

important concerns with EBA that include, among others, insensitivity to the relative goodness

of fit of different models, so that a misspecified model may determine the robustness of a

variable. In addition, they indicate that EBA requires an excessively strong standard for

robustness. As recognized by Gassebner and Luechinger (2011), their modified EBA requires

linear estimates. Moreover, their exercise cannot account for structural relations, omitted

variable biases, or simultaneity concerns. The linearity issue is particularly germane for the

alleged determinants of terrorism because the next two subsections indicate that nonlinear

estimations best capture the relationship between GDP per capita and terrorism and the

relationship between regime type and terrorism.

In summary, there is little convincing evidence supporting globalization as a cause of

transnational terrorism; however, there is room for a better methodology to address this cause.

For the present, we must look elsewhere for a cause.

6.2 Poverty and Transnational Terrorism

An initial influential study on the potential relationship between poverty and terrorism is by

Krueger and Malečková (2003), who first present evidence from Palestinian opinion polls that

show that a large majority of highly educated, and presumably better off, Palestinians supported

43 See Drakos and Gofas (2006), Kurrild-Klitgaard, Justesen, and Klemmensen (2006), and Gaibulloev, Piazza, and Sandler (2017).

54

the use of terrorism. Next, these authors examine the biographical information of 129 members

of Hezbollah’s military wing, whose 28% poverty rate was five percent lower than that of the

host population. Moreover, these fighters’ education attainment was statistically greater than

that of the general public. Their final bit of evidence comes from a negative binomial cross-

country analysis of 129 to 148 countries during 1997–2002, for which they find a marginal

negative relationship (p = 0.11) between transnational terrorism and GDP per capita. However,

when they control for civil liberties, the poverty link vanishes, leading Krueger and Malečková

(2003) to conclude that lack of civil liberties, and not poverty, drives transnational terrorism.

This finding is similar in spirit to that of Krueger and Laitin (2008) who demonstrate that

political repression, not GDP per capita, correlated with transnational terrorist incidents in the

perpetrators’ home countries.

The literature offers a wide variety of results on the relationship between GDP per capita

and terrorism. Some articles display a positive relationship, (e.g., Bandyopadhyay and Younas

2011, Blomberg, Hess, and Weerapana 2004, and Gassebner and Luechinger 2011); other

articles uncover a negative relationship (e.g., Li and Schaub 2004); and still others find no

relationship (e.g., Abadie 2006 and Krueger and Malečková 2003). Most articles on this

empirical relationship rely on panel estimates and pay little attention to functional form. A few

studies uncover an inverted U-shaped relationship between terrorism and GDP per capita (e.g.,

Enders and Hoover 2012 and Lai 2007).

The literature uses different country samples, diverse time periods, and alternative

measures of transnational terrorism. Hence, a robust methodology is needed that can produce

consistent findings for varying controls, sample periods, and sample countries. To date, the most

robust analysis is that of Enders, Hoover, and Sandler (2016) who employ a flexible nonlinear

55

empirical specification that includes linear, quadratic and other functional forms.44 Their

empirical analysis is performed on alternative series of terrorist incidents that involve casualties,

since such incident are more apt to be reported and less driven by random factors. Their

regressions indicate a consistent nonlinear relationship between real GDP per capita and eight

alternative measures of terrorism, accounting for attack venue, perpetrators’ nationality, and

terrorism type (i.e., transnational and domestic). Two alternative time periods – 1970–1992 and

1994–2010 – are analyzed that correspond to the dominance of the leftist and religious

fundamentalist terrorists, respectively. In all instances, their nonlinear estimates outperform

linear and quadratic representations.

For all eight terrorism series, there is a peak level of terrorism at some intermediate range

of GDP per capita. There is a nonsymmetrical relationship on either side of the peak, leading to

a rejection of a quadratic relationship between income per capita and terrorism. This peak

shifted between the two time periods: occurring at a relatively high intermediate GDP per capita

when the leftist terrorists dominated in the pre-1993 era, and occurring at a relatively low

intermediate GDP per capita when the religious fundamentalist terrorists dominated in the post-

1993 era. This shifting peak is consistent with many left-wing terrorist groups originating from

and operating in relatively well-to-do countries during the earlier period, and many religious

fundamentalist terrorist groups originating from and operating in less well-to-do countries during

the latter period. The former consideration may explain the preponderance of positive linear

estimates for the income-terrorism relationship before the early 1990s. The Enders, Hoover, and

Sandler (2016) study shows that neither very low nor very high GDP per capita is conducive to

terrorism. At low GDP per capita, subsistence is an overriding concern, thus limiting terrorism.

At high GDP per capita, grievances are not great and society invests in counterterrorism, thus

44 Their method is known as a nonlinear smooth transition regression.

56

curbing terrorism. Moreover, these nonlinear relationships are robust to myriad controls that

display anticipated signs.45 The relationship between GDP per capita and terrorism is much

more nuanced than presupposed by political leaders, the media, and researchers. Not only is the

relationship nonlinear, but its nature alters with circumstances – e.g., the predominant type of

terrorism and the origin country of perpetrators – thereby confounding estimation. In summary,

there is no discernible relationship per se between poverty and alternative forms of terrorism.

6.3 Regime Type and Transnational Terrorism

Given that 9/11 took place in a democracy and high-profile transnational terrorist attacks before

and after 9/11 targeted democracies, the relationship between regime type and terrorism is of

great interest. There are three schools of thought on this relationship. The “strategic” school

argues that democracy facilitates terrorism by limiting its price through freedom of movement,

freedom of association, the protection of civil liberties, rights to due process, and the ability to

acquire weapons (Eubank and Weinberg 1994, Eyerman 1998, and Hoffman 2006). Moreover,

in democracies, freedom of the press provides terrorists with the publicity that they crave (Li

2005 and Rohner and Frey 2007). By contrast, the “political access” school views democracies

as assuaging grievances through greater political participation, thereby reducing terrorism

(Eyerman 1998 and Li 2005). The “democratic protection” school stresses that strong liberal

democracies take swift actions to counter terrorism in order to maintain their legitimacy

(Gaibulloev, Piazza, and Sandler 2017). Strict autocracies also take draconian measures to

counter terrorism. Given these opposing considerations, the relationship between regime type

and transnational terrorism is an empirical question.

Findings in the empirical literature on this relationship are mixed and generally

45 Domestic terrorism displays the same nonlinear relationship with GDP per capita.

57

unconvincing. Most articles find a positive relationship between democracy and terrorism in

support of the strategic school (e.g., Chenoweth 2010, Eubank and Weinberg 1994, 2001, Lai

2007, Li and Schaub 2004, Piazza 2008, and San-Akca 2014). A small number of articles

display a negative relationship, supportive of the political access or democratic protection

schools (e.g., Eyerman 1998). Other articles find no significant relationship (Gassebner and

Luechinger 2011, Krueger and Laitin 2008, and Savun and Phillips 2009).46 A handful of

articles uncover an inverted U-shaped relationship in which some intermediate regime type in

between autocracy and democracy, known as anocracy, displayed the greatest amount of

terrorism (Abadie 2006, Chenoweth 2013, Drakos and Gofas 2006, and Gaibulloev, Piazza, and

Sandler 2017).

The overwhelming number of studies involves panel analyses that do not adequately

address endogeneity concerns regarding regime type and terrorism. Many of these studies either

do not account for unobserved heterogeneities or, when they do, their fixed-effects estimates

typically do not support the articles’ hypotheses. Often, articles examine limited time periods

and a single type of terrorism, leaving one to wonder about the true relationship between regime

type and alternative forms of terrorism.

Unlike the extant literature, Gaibulloev, Piazza, and Sandler (2017) take up these

concerns in their investigation of regime type and terrorism for a global sample during 1970–

2012. They first present a game-theoretic model, accounting for the three schools of thought. In

so doing, they establish theoretic grounds for an inverted U-shaped relationship between forms

of terrorism – transnational and domestic – and alternative measures of regime type.47 This

follows because autocracies employ draconian measures to curb terrorism, while strong 46 Sandler (1995) cautions about the bias from employing terrorist event datasets, collected from the news media, to test this relationship. Autocracies are less likely than democracies, to report small-scale terrorist incidents. There are ways to ameliorate this bias. 47 Also see Dragu (2011) on an inverted U-shaped relationship between security and liberty.

58

democracies also apply effective means to protect lives and property. Moreover, strong

democracies limit terrorism-inducing grievances through political access. In contrast, anocracies

lack the means to respond rigorously to terrorism and only offer intermediate levels of political

access. Along a regime spectrum with autocracies and strong democracies at the endpoints,

anocracies are anticipated to experience the most domestic and transnational terrorism, thus

giving rise to an inverted U-shaped regime-terrorism relationship.

Given the count nature of the terrorism variable, Gaibulloev, Piazza, and Sandler (2017)

employ various estimators for panel count data to test empirically their hypothesis. They

consider an extensive set of controls, consistent with the literature, and present instrumental

variable approaches to address endogeneity concerns. Alternative measures of regime type are

included to capture notions of political rights, civil rights, and political participation. These

authors show that regime type has a robust inverted U-shaped relationship on terrorism. Their

results strongly suggest that the effect of regime type on terrorism is nonlinear and more

complex than initially presupposed by commentators, leaders, and researchers.

This inverted U-shaped finding raises an important policy concern with respect to post-

9/11 efforts to push democracies, culminating in Western democracies supporting the “Arab

Spring.” As autocracies are transformed into fledgling democracies or anocracies, the risk of

greater terrorism surfaces, so that preventive actions must be in place to protect against this risk.

7. Economic Consequences of Terrorism

Terrorists aim to harm the economy of targeted governments in order to induce governments’

concessions to some of the terrorists’ demands. But how damaging is terrorism to a country’s

economy?

59

7.1 Economic Growth and Terrorism

Based on a panel of 177 countries for 1968–2000, Blomberg, Hess, and Orphanides (2004)

uncover a negative and statistically significant effect of transnational terrorism on the growth of

GDP per capita. Terrorism appears to induce a reallocation of resources from productive

investments toward government spending. However, the magnitude of the growth effect of

terrorism is economically modest and substantially smaller than the impact of internal or external

conflicts.48 Their analysis shows that the impact of terrorism on GDP disappears after one year,

which is much shorter-lived relative to the effect of internal and external conflicts. Using a panel

of world countries for 1987–2001, Tavares (2004) also finds that the negative effect of

transnational terrorism on economic growth is quantitatively small. Moreover, the estimated

coefficient for terrorism is not statistically robust across alternative model specifications.

Subsequent studies distinguish between transnational and domestic terrorism and focus

on regional subsamples (e.g., Africa, Asia, and Western Europe).49 Generally, terrorism had a

small adverse effect on the growth of GDP per capita. The effect is driven by transnational

terrorism; domestic terrorism is usually not statistically significant.50

This literature commonly applies dynamic fixed-effects regressions to account for

endogeneity bias that might stem from country-specific unobserved heterogeneities and time-

specific global shocks. However, the issue of endogeneity of terrorism – owing to potential

omission of factors that may influence both terrorism and economic growth – remains a concern.

Given the difficulty of finding a proper instrument, some studies use lagged value of terrorism to

48 A one-unit increase in the number of terrorist attacks per million persons was associated with 0.24 percentage point decrease in GDP per capita growth for the full-sample fixed-effects regression (Blomberg, Hess, and Orphanides 2004, Table 5, p. 1023). The average number of terrorist attacks per million persons was just 0.2 in their sample. 49 See, e.g., Gaibulloev and Sandler (2008, 2009b, 2011). 50 Findings also suggest that less-developed countries were more susceptible to a negative impact of terrorism on growth (e.g., Blomberg, Hess, and Orphanides 2004).

60

ameliorate this issue (e.g., Tavares 2004). Using fixed-effects estimator in dynamic panel

regression leads to Nickel bias. To address this issue, studies focus on longer sample periods and

relatively smaller number of cross-sectional units. In a related study, Gaibulloev, Sandler, and

Sul (2013) find a significant cross-sectional correlation of transnational terrorism across 106

countries. Gaibulloev, Sandler, and Sul (2014) show that the negative, albeit small, impact of

terrorism on growth is not robust: once Nickell bias and cross-sectional dependence are corrected

in dynamic panel regressions, the coefficient for terrorism loses its statistical significance.

With few exceptions, the literature does not distinguish between the pre-1990s and the

post-1990s periods in terms of the effect of terrorism on growth. For Asia, Gaibulloev and

Sandler (2009b) find that terrorism is not statistically significant for the period before 1986 but is

statistically significant and negative for 1986–2004.

The adverse effect of terrorism on growth can be substantial for an individual country or

subregion plagued with terrorist incidents. Abadie and Gardeazabal (2003) introduce an

innovative econometric method – synthetic control approach – to obtain the causal effect of

terrorism on the economy of the Basque Country.51 This method addresses identification

challenges that arise in cross-country panel regressions. They show that terrorism reduced GDP

per capita in the Basque Country by about 10 percentage points during the 1980s and 1990s.

These authors further demonstrate that the 1998–1999 cease-fire caused a substantial

improvement in the relative performance of the stocks portfolio of firms with large business

dealings in the Basque region (value increase of 10.4%) when the truce became credible. The

abnormal return of Basque stocks relative to non-Basque stocks was negative and large (–11.2%)

51 The basic idea is to construct a “synthetic” Basque Country using a weighted combination of the other regions in Spain so that the “synthetic” and real Basque Country are closely matched in terms of their economic characteristics before the onset of terrorism. Assuming that the economic trajectories of “synthetic” Basque Country and the Basque Country are similar in the absence of terrorism, any differences in the outcome variable after the outbreak of terrorism in the Basque Country is attributed to the casual effect of terrorism.

61

at the end of the truce period. A case study of Israel also indicates a sizable impact of terrorism

on per capita output during high terrorism periods (Eckstein and Tsiddon 2004).

7.2 Trade and Terrorism

In general, the empirical research on terrorism and trade finds a negative relationship between

terrorism and bilateral trade.52 Presumably, trade falls owing to greater cost of doing business –

e.g., greater uncertainty, higher insurance and wage premiums, and larger transaction costs. The

literature usually utilizes the gravity model, which is the standard in the empirical trade

literature. The adverse effect of terrorism on trade is shown for the period of the dominance by

the leftist terrorist groups (e.g., Nitsch and Schumacher 2004) as well as for the era of the

dominance by the religious fundamentalist terrorists (e.g., Bandyopadhyay, Sandler, and Younas

2018). Egger and Gassebner (2015) is an exception; they discern a negligible nonrobust effect of

terrorism on trade in the short run. They attribute the differences in their findings to the neglect

of aggregation bias and general equilibrium effects by previous studies. Although most work on

terrorism and trade use transnational terrorism data, recent evidence suggests that domestic

terrorism also reduced bilateral trade. For example, Bandyopadhyay, Sandler, and Younas

(2018) disaggregate bilateral trade data into primary commodities and manufactured goods for a

panel of 151 countries during 1995–2012 and find a detrimental impact of both domestic and

transnational terrorism on trade, particularly on the trade of manufactured goods.53 Overall,

terrorism had a considerable smaller effect on trade relative to other types of violence, such as

52 See, e.g., Blomberg and Hess (2006), De Sousa, Mirza, and Verdier (2009), and Mirza and Verdier (2014). These studies apply gravity trade models using panel data. De Sousa, Mirza, and Verdier (2009) examine the spatial spillover effect of terrorism and find that a country’s export to the United States was adversely affected by terrorism in neighboring countries. 53 Nitsch and Schumacher (2004) estimate the elasticity of total trade in response to transnational terrorism to be around –0.04 using pre-2001 data. Blomberg and Hess (2006) use a dummy variable for terrorism: if a country-pair experienced terrorism in a given year, trade fell by about 5%.

62

civil war (e.g., Blomberg and Hess 2006).

7.3 Stock Exchanges and Terrorism

Chen and Siems (2004) implement the event-study methodology to examine the effect of

terrorist attacks on global capital markets. An analysis of 14 military/terrorist events for 1915–

2001 reveals that, over time, the US capital markets became more resilient and faster in

absorbing the effects of such shocks.54 US markets displayed a remarkable resilience to the 9/11

terrorist attacks. The decline in the US stock market, though significant, was the second smallest

among the 10 largest capital markets in the world. The Dow Jones Industrial Average stock

index took 40 trading days to fully recover its losses.

Chen and Siems’ (2004) analysis illustrates that a large terrorist attack could have a

sizable effect on financial markets, but the impact was transitory; markets absorbed the shock

and continued to function efficiently. Within 60 trading days of 9/11, 27 of the 33 markets

(82%) in their sample recouped their losses. The length and the depth of the terrorism shock

depended on the size and the maturity of the markets and the strength of institutions. Stable

banking/financial sector and prudent government policies can mitigate the adverse effect of

terrorism by providing necessary liquidity (and fiscal stimulus) as shown in the case of the

United States (Enders and Sandler 2012). This also implies that developing markets are more

vulnerable to terrorism. In fact, Arin, Ciferri, and Spagnolo (2008) compare six financial

markets and record a greater effect of terrorism on stock returns and stock volatility in emerging

markets.

Terrorism has a dissimilar effect on stocks both across countries and industries. Chesney, 54 Out of eight terrorist attacks in the sample, only the 9/11 attacks had a significant impact on the Dow Jones Industrial average stock index. The Air India bombing in 1985 had a marginally significant impact only on the day of the event. The other terrorist incidents, which include the US Embassy bombing in Kenya in 1998 and the World Trade Center bombing in 1993, did not have a statistically significant effect.

63

Reshetar, and Karaman (2011) employ various empirical methodologies, including event-study

and non-parametric approaches, to compare the effects of 77 major terrorist events, 4 financial

crashes, and 19 natural disasters for 1994–2005 on global and regional stock markets and

industry stock indices. About two-thirds of the sample terrorist attacks had a significant adverse

impact on at least one of the studied stock markets. The American (Swiss) market is found to be

most (least) immune to the negative shock of terrorism. Similarly, the majority of the sample

terrorist events had a significant impact on at least one of the studied industry indices. However,

these authors do not examine why a given stock market or industry stock index reacted

differently to individual terrorist events. Chesney, Reshetar, and Karaman (2011) uncover

interesting similarities and differences between three types of shocks – terrorism, financial crises,

and natural disasters – in terms of their effect on industry indices. For example, terrorism had a

strong adverse effect on the stocks of airline.55 In contrast, the banking sector was most

vulnerable to financial crises. The impact of terrorism, similar to financial crises, was strongest

during an event-day but decayed in the post-event period.

7.4 Direct and Indirect Costs of Large-Scale Terrorist Incidents

Available estimates suggest that the direct economic cost of large-scale terrorist incidents is

modest, particularly for advanced economies. According to Kunreuther, Michel-Kerjan, and

Porter (2003) the damage from the 9/11 terrorist attacks was more than $80 billion for the United

States. Similarly, the estimated direct cost of the March 11, 2004 terrorist attacks in Madrid,

Spain, was around 212 million euros, which was about 0.03% of Spain’s GDP (Buesa et al.

2007). As pointed out by Becker and Murphy (2001), diversified economies adjust and

55 Drakos (2004) discerns a substantial increase in both systematic and idiosyncratic risks of the airline stocks as a result of 9/11.

64

withstand large-scale attacks with a small number of key sectors affected. They estimate the cost

of the ongoing threat following 9/11 to be around 0.3% of the US GDP.56 However, indirect

economic costs of terrorist attacks, which are difficult to measure, are likely to be substantial.57

7.5 Tourism and Terrorism

The tourism sector is particularly vulnerable to terrorism. Enders, Sandler, and Parise (1992)

examine the impact of terrorist events on tourism receipts in Western European nations using

quarterly data for 1970–1988. Using an ARIMA method with transfer function, these authors

find that terrorism reduced tourism revenue. However, the effect was heterogeneous: Austria,

Greece, and Italy suffered significant losses in tourism receipts but France, Denmark, and

Germany did not experience statistically significant losses. Further, there was evidence for a

negative externality; terrorist incidents in a country dissuaded tourist visitors in neighboring

countries. A number of ensuing studies examine an individual country, a region, and a world

sample for either the 1970s–2000s or the post-1990s period; they confirm that terrorism

adversely influences tourism.58 Overall, the short-run effect of terrorism on tourism can be large

but the impact faded quickly with airline discount fares. Nevertheless, for countries that rely on

tourism for a source of foreign exchange, sustained terrorist campaigns can be detrimental to the

economy.

Recent studies point toward a significant spillover effect from terrorism. Drakos and

Kutan (2003) uncover a negative association between terrorist incidents and the number of

tourists that used three Mediterranean tourist destinations – Greece, Israel, and Turkey – from

56 They assume one successful attack annually that downs one plane and kills up to 100 people. The estimated cost includes air travel security, loss of human life, physical damage, and reduced GDP growth. 57 See Frey, Luechinger, and Stutzer (2009) for estimates of the psychological cost of terrorism. 58 See, e.g., Neumayer (2004), Pizam and Fleischer (2002), and Sloboda (2003). Ito and Lee (2005) report a significant immediate drop of 30% in demand for US airline as a result of 9/11 attacks.

65

January 1991 to December 2000. As the intensity of attacks increased in a country, the number

of tourists fell, causing the country to lose market share. The tourists either chose another

country in the region (substitution effect) or left the region in favor of safer destinations (regional

effect). The latter effect is estimated to be substantially greater than the substitution effect for

the sample countries. Neumayer and Plümper (2016) distinguish three types of statistically

significant spillover effects arising from transnational terrorist campaigns in Islamic countries. 59

A terrorist attack against a Westerner in an Islamic country deterred the flow of tourists from the

victim’s country to other Islamic countries. Furthermore, the attack discouraged tourists from

other Western countries to both the targeted Islamic country and other Islamic destination

nations.

7.6 Foreign Direct Investment and Terrorism

International investors consider countries’ risk ratings, including terrorist risk, in their

investment decisions. Terrorist risk reduces the expected return to investment and creates

uncertainty about possible future terrorist incidents (Abadie and Gardeazabal 2008). Therefore,

in an open economy, mobility of productive resources results in the reallocation of investments

from countries with higher level of terrorism to those with small levels of attacks. For a sample

of 78 developing countries over 1984–2008, Bandyopadhyay, Sandler, and Younas (2014)

uncover a negative association between both transnational and domestic terrorist incidents and

foreign direct investment (FDI). The specification challenges discussed in Section 7.1 apply to

FDI regressions as well. Bandyopadhyay, Sandler, and Younas (2014) implement Dynamic

GMM method to address the Nickel bias and endogeneity problems. The majority of studies use

59 They use a dyadic data of world countries for 1995–2013. The regression model is specified following a gravity-type model, which is well-suited to explain the dyadic relationship.

66

terrorist incidents and casualties; however, Abadie and Gardeazabal (2008) take advantage of a

unique data on terrorist risk across 186 countries for 2003/2004. Their cross-sectional analysis

reveals that the net FDI position of an average country fell by about 5% of GDP in response to a

one standard deviation increase in the terrorist risk, which is a large shock. These authors

provide numerous robustness checks to eliminate potential sources of the omitted variable bias

and argue that a potential reverse causation is not expected to weaken their result.

In general, the available evidence suggests that terrorism depressed FDI both before 1990

(e.g., Enders and Sandler 1996) and after 1990, but the magnitude of the effect depends on the

size of the economy and the severity of terrorism.60 Generally, the effect is small in relation to

the size of the economy, especially because an average country experiences very little

transnational terrorism.

8. Concluding Remarks

To conclude this study, we return to address the questions posed at the outset.

Over the last half century, transnational and domestic terrorism changed in a few

essential ways. First, the number of transnational terrorist incidents fell by about 40% since the

start of the 1990s; however, each incident was much more likely to involve casualties since then.

Second, domestic terrorism displayed a somewhat similar pattern: it rose until 1992 and then fell

before rising abruptly after 2008. After 9/11, transnational and domestic terrorists shifted their

venue from Europe, Central Asia, and Latin America to the Middle East, Africa, and South Asia.

Some of this displacement is due to defensive measures taken by industrial countries following

60 Using quarterly data for 1975–1991, Enders and Sandler (1996) find that the annual number of terrorist incidents in Spain (Greece) reduced Spanish (Greek) annual net FDI by 13.5% (11.9%). A panel analysis of 69 countries for 1989–1999 suggests that attacks against US interests abroad had a small effect on the stock of US FDI in the OECD countries but had no effect on the stock of US FDI in the non-OECD countries (Enders, Sachsida, and Sandler 2006).

67

9/11. Third, during the post-9/11 era, terrorists relied more heavily on hostage taking and armed

attacks. The former provided funding that was reduced by post-9/11 actions.

Religious fundamentalist groups, which represent the dominant terrorist influence for the

last 25 years, rely on kinship, long-term friendships, and worship for recruitment. Some of these

groups tighten their grip on members by supplying excludable club goods to operatives (Berman

2009). Backlash to governments’ crackdowns can also attract recruits, as can highly visible

spectacular terrorist attacks, even when unsuccessful. The literature on terrorist group shows

that religious fundamentalist groups display long-term survival, fostered by group size and

cooperative links with other terrorist groups. Islamic terrorist groups are particularly resilient.

Groups that rely less on transnational terrorist attacks possess better survival prospects by not

angering a formidable targeted state. Terrorist groups foster their survivability by responding

rationally to government countermeasures.

Studies of counterterrorism effectiveness offer a rather mixed picture. Technological

barriers are effective provided that they are constantly updated, so that they stay ahead of

terrorist-designed circumventions. In addition, authorities must anticipate attack, target, and

location transference that barriers and other defenses encourage. For transnational terrorism,

defensive measures tend to be overprovided, while proactive measures tend to be underprovided,

so that targeted governments work at cross-purposes with one another. Thus, international

cooperation has the potential of achieving the right balance of these measures against a common

terrorist threat by internalizing a host of opposing externalities. Unfortunately, even after 9/11,

the extent of such cooperation is rather limited as targeted countries cherish their autonomy over

security matters. At the domestic level, the right mix of defensive and proactive measures is

generally struck as a central government internalizes externalities between alternative venues and

targets. Berman (2009) offers a novel counterterrorism approach against Islamic fundamentalist

68

terrorists that involves government provision of services to counter monopoly provision of these

services by some religious fundamentalist groups. The literature also shows that directed

proactive measures – e.g., assassination of militant leaders or house demolitions – are effective

(Benmelech, Berrebi, and Klor 2015 and Zussman and Zussman 2006).

Following 9/11, the media, some heads of state, and academics put forward three causes

or facilitators of terrorism: globalization, poverty, and democracy. To date, the literature finds

no evidence of a relationship between terrorism and globalization. A similar story holds for

poverty. Any relationship between GDP per capita and terrorism is highly nonlinear with some

intermediate level of income being most associated with transnational and domestic terrorism.

Operationally, terrorist leaders prefer to deploy well-educated operatives, even for suicide

missions, because they are likely to succeed (Benmelech and Berrebi 2007). Moreover,

anocracies, not democracies, provide the most favorable environment for terrorism because

political access is limited, many terrorism-facilitating freedoms are present, and lives and

property are not well-protected. The alleged relationships between GDP per capita and terrorism

and between regime type and terrorism are more subtle than originally presupposed.

Despite the damage wreaked by 9/11, most terrorist attacks have little impact on property

and people on average. For large-scale terrorist attacks, economic and financial ramifications are

transitory. Most countries sustain very few terrorist attacks annually so that terrorism generally

has little or no influence on GDP per capita growth. Rich diversified countries are able to absorb

these attacks with little macroeconomic consequences; however, terrorism-plagued small or

developing countries are adversely affected in terms of lost GDP. Losses in terrorism-prone

sectors (e.g., tourism and FDI) may be measurable but relatively small compared to GDP for

most countries. Rich diversified economies naturally cushion these sectoral impacts by

transference as activities move to less terrorism-risky parts of a targeted country’s economy.

69

There are clear directions for future research on terrorism. Empirical analyses on aspects

of terrorism must pay greater heed to their identification strategy. Panel studies on terrorism

determinants need to do more than lag some control variables to account for endogeneity. Given

the rise of homegrown terrorist attacks by Islamic extremist groups, research must consider the

radicalization process and subsequent attacks. To assist this research, terrorist event datasets

should record whether attacks are inspired by foreign influences and radicalization. Currently,

group level data that are cleaned and tied to an event dataset ends in 2006. Such data must be

brought up-to-date to investigate how diverse terrorist ideologies influence groups’ survival and

their response to counterterrorism. In particular, this expanded data on terrorist groups will

foster more informative studies on the determinants of alternative ways that these groups end in

the post-9/11 era. Attention must be paid to how micro-level studies of particular terrorist

campaigns add to our general knowledge of terrorism. For instance, do micro-level studies of

Israel or Iraq inform us about terrorist campaigns in other places where different conditions

apply? Also, the relationship between immigration and terrorism warrants careful study.

Finally, empirical analyses of terrorist behavior must account for more competing choices.

70

References

Abadie, Alberto. 2006. “Poverty, Political Freedom and the Roots of Terrorism.” American

Economic Review 96 (2): 50–6.

Abadie, Alberto and Javier Gardeazabal. 2003. “The Economic Cost of Conflict: A Case Study

of the Basque Country.” American Economic Review 93 (1): 113–32.

Abadie, Alberto and Javier Gardeazabal. 2008. “Terrorism and the World Economy.”

European Economic Review 52 (1): 1–27.

Abrahms, Max. 2006. “Why Terrorism Does Not Work.” International Security 31 (2): 42–78.

Abrahms, Max. 2008. “What Terrorists Really Want: Terrorist Motives and Counterterrorism

Strategy.” International Security 32 (4): 78–105.

Anderton, Charles H. and John R. Carter. 2005. “On Rational Choice Theory and the Study of

Terrorism.” Defence and Peace Economics 16 (4): 275–82.

Arce, Daniel G., Rachel T. A. Croson, and Catherine C. Eckel. 2011. “Terrorism Experiments.”

Journal of Peace Research 48 (3): 373–82.

Arce, Daniel G. and Todd Sandler. 2005. “Counterterrorism: A Game-Theoretic Analysis.”

Journal of Conflict Resolution 49 (2): 183–200.

Arce, Daniel G. and Todd Sandler. 2007. “Terrorist Signalling and the Value of Intelligence.”

British Journal of Political Science 37 (4): 573–86.

Arce, Daniel G. and Todd Sandler. 2010. “Terrorist Spectaculars: Backlash Attacks and the

Focus of Intelligence.” Journal of Conflict Resolution 54 (2): 354–73.

Arce, Daniel G. and Kevin Siqueira. 2014. “Motivating Operatives for Suicide Missions and

Conventional Terrorist Attacks.” Journal of Theoretical Politics 26 (4): 677–95.

Arin, Peren K., Davide Ciferri, and Nicola Spagnolo. 2008. “The Price of Terror: The Effects

of Terrorism on Stock Market Returns and Volatility.” Economics Letters 101 (3): 164–

71

7.

Asworth, Scott, Joshua D. Clinton, Adam Meirowitz, and Kristopher W. Ramsey. 2008.

“Design, Inference, and the Strategic Logic of Suicide Terrorism.” American Political

Science Review 102 (2): 267–72.

Atkinson, Scott E., Todd Sandler, and John Tschirhart. 1987. “Terrorism in a Bargaining

Framework.” Journal of Law and Economics 30 (1): 1–21.

Azam, Jean-Paul. 2005. “Suicide-Bombing as Inter-Generational Investment.” Public Choice

122 (1–2): 177–98.

Azam, Jean-Paul and Alexandra Delacroix. 2006. “Aid and the Delegated Fight against

Terrorism.” Review of Development Economics 10 (2): 330–44.

Azam, Jean-Paul and Véronique Thelen. 2008. “The Roles of Foreign Aid and Education in the

War on Terror.” Public Choice 135 (3–4): 375–97.

Azam, Jean-Paul and Véronique Thelen. 2010. “Foreign Aid versus Military Intervention in the

War on Terror.” Journal of Conflict Resolution 54 (2): 237–61.

Bandyopadhyay, Subhayu and Todd Sandler. 2011. “The Interplay between Preemptive and

Defensive Counterterrorism Measures: A Two-Stage Game.” Economica 78 (3): 546–

64.

Bandyopadhyay, Subhayu and Todd Sandler. 2014. “Immigration Policy and

Counterterrorism.” Journal of Public Economics 110: 112–23.

Bandyopadhyay, Subhayu, Todd Sandler, and Javed Younas. 2011. “Foreign Aid as

Counterterrorism Policy.” Oxford Economic Papers 63 (3): 423–47.

Bandyopadhyay, Subhayu, Todd Sandler, and Javed Younas. 2014. “Foreign Direct Investment,

Aid, and Terrorism.” Oxford Economic Papers 66 (1): 25–50.

Bandyopadhyay, Subhayu, Todd Sandler, and Javed Younas. 2018. “Trade and Terrorism: A

72

Disaggregated Approach.” Journal of Peace Research 55 ( ): DOI:

10.1177/0022343318763009, forthcoming..

Bandyopadhyay, Subhayu and Javed Younas. 2011. “Poverty, Political Freedom, and the Roots

of Terrorism in Developing Countries: An Empirical Assessment.” Economics Letters

112 (2): 171–76.

Bapat, Navin A. 2007. “The Internationalization of Terrorist Campaigns.” Conflict

Management and Peace Science 24 (4): 265–80.

Bapat, Navin A. 2011. “Transnational Terrorism, US Military Aid, and the Incentive to

Misrepresent.” Journal of Peace Research 48 (3): 303–18.

Basile, Mark. 2004. “Going to the Source: Why Al Qaida’s Financial Network is Likely to

Withstand the Current War on Terrorist Financing.” Studies in Conflict & Terrorism 27

(3): 169–85.

Becker, Gary S. and Kevin M. Murphy. 2001. “Prosperity Will Rise Out of the Ashes?” Wall

Street Journal (Eastern ed.), 29 October 2001: A.22.

Benmelech, Efraim and Claude Berrebi. 2007. “Human Capital and the Productivity of Suicide

Bombers.” Journal of Economic Perspectives 21 (2): 223–38.

Benmelech, Efraim, Claude Berrebi, and Esteban F. Klor. 2012. “Economic Conditions and the

Quality of Suicide Terrorism.” Journal of Politics 74 (1): 113–28.

Benmelech, Efraim, Claude Berrebi, and Esteban F. Klor. 2015. “Counter-Suicide-Terrorism:

Evidence from House Demolitions.” Journal of Politics 77 (1): 27–43.

Berman, Eli. 2000. “Sect, Subsidy and Sacrifice: An Economist’s View of Ultra-Orthodox

Jews.” Quarterly Journal of Economics 115 (3): 905–53.

Berman, Eli. 2009. Radical, Religious, and Violent: The New Economics of Terrorism.

Cambridge, MA: The MIT Press.

73

Berman, Eli and David D. Laitin. 2008. “Religion, Terrorism, and Public Clubs: Testing the

Club Model.” Journal of Public Economics 92 (9–10): 1942–67.

Berman, Eli, Jacob N. Shapiro, and Joseph H. Felter. 2011. “Can Hearts and Minds Be Bought?

The Economics of Counterinsurgency in Iraq.” Journal of Political Economy 119 (4):

766–819.

Berrebi, Claude and Esteban F. Klor. 2006. “On Terrorism and Electoral Outcome: Theory and

Evidence from the Israeli-Palestinian Conflict.” Journal of Conflict Resolution 50 (6):

899–925.

Berrebi, Claude and Darius Lakdawalla. 2007. “How Does Terrorism Risk Vary Across Space

and Time? An Analysis Based on Israeli Experience.” Defence and Peace Economics 18

(2): 113–31.

Bier, Vicki, Santiago Oliveros, and Larry Samuelson. 2007. “Choosing What to Protect:

Strategic Defensive Allocation against an Unknown Attacker.” Journal of Public

Economic Theory 9 (4): 563–87.

Blomberg, S. Brock, Rozlyn C. Engel, and Reid Sawyer. 2010. “On the Duration and

Sustainability of Transnational Terrorist Organizations.” Journal of Conflict Resolution

54 (2): 303–30.

Blomberg, S. Brock, Khusrav Gaibulloev, and Todd Sandler. 2011. “Terrorist Group Survival:

Ideology, Tactics, and Base of Operations.” Public Choice 149 (3–4): 441–63.

Blomberg, S. Brock and Gregory D. Hess. 2006. “How Much Does Violence Tax Trade?”

Review of Economics and Statistics 88 (4): 599–612.

Blomberg, S. Brock, Gregory D. Hess, and Athanasios Orphanides. 2004. “The Macroeconomic

Consequences of Terrorism.” Journal of Monetary Economics 51 (5): 1007–32.

Blomberg, S. Brock, Gregory D. Hess, and Akila Weerapana. 2004. “Economic Conditions and

74

Terrorism.” European Journal of Political Economy 20 (2): 463–78.

Bloom, Mia. 2005. Dying to Kill: The Allure of Suicide Terror. New York: Columbia

University Press.

Brandt, Patrick T., Justin George, and Todd Sandler. 2016. “Why Concessions Should Not Be

Made to Terrorist Kidnappers.” European Journal of Political Economy 44: 41–52.

Brandt, Patrick T. and Todd Sandler. 2009. “Hostage Taking: Understanding Terrorism Event

Dynamics.” Journal of Policy Modeling 31 (5): 758–78.

Brandt, Patrick T. and Todd Sandler. 2010. “What Do Transnational Terrorists Target? Has It

Changed? Are We Safer?” Journal of Conflict Resolution 54 (2): 214–36.

Brandt, Patrick T. and Todd Sandler. 2012. “A Bayesian Poisson Vector Autoregression

Model.” Political Analysis 20 (3): 292–315.

Brock, William A., Steven N. Durlauf, and Kenneth D. West. 2003. “Policy Evaluation in

Uncertain Economic Environments.” Brookings Papers on Economic Activity 2003 (1):

235–301.

Brophy-Baermann, Bryan and John A. C. Conybeare. 1994. “Retaliating against Terrorism:

Rational Expectations and the Optimality of Rules versus Discretion.” American Journal

of Political Science 38 (1): 196–210.

Bueno de Mesquita, Ethan. 2005a. “Conciliation, Commitment and Counterterrorism: A

Formal Model.” International Organization 59 (1): 145–76.

Bueno de Mesquita, Ethan. 2005b. “The Quality of Terror.” American Journal of Political

Science 49 (3): 515–30.

Bueno de Mesquita, Ethan. 2007. “Politics and the Suboptimal Provision of Counterterror.”

International Organization 61 (1): 9–36.

Bueno de Mesquita, Ethan. 2013. “Rebel Tactics.” Journal of Political Economy 121 (2): 323–

75

57.

Buesa, Mikel, Aurelia Valiño, Joost Heijs, Thomas Baumert, and Javier González Gómez. 2007.

“The Economic Cost of March 11: Measuring Direct Economic Cost of the Terrorist

Attack on March 11, 2004 in Madrid.” Terrorism and Political Violence 19 (4): 489–

509.

Bush, George W. 2002. Securing the Homeland: Strengthening the Nation. Washington, DC:

White House.

Cárceles-Poveda, Eva and Yair Tauman. 2011. “A Strategic Analysis of the War against

Transnational Terrorism.” Games and Economic Behavior 74 (1): 49–65.

Carter, David B. 2010. “The Strategy of Territorial Conflict.” American Journal of Political

Science 54 (4): 969–87.

Carter, David B. 2012. “A Blessing or Curse? State Support for Terrorist Groups.”

International Organization 66 (1): 129–51.

Carter, David B. 2015a. “When Terrorism Is Evidence of State Success: Securing the State

against Territorial Groups.” Oxford Economic Papers 67 (1): 116–32.

Carter, David B. 2015b. “The Compellence Dilemma: International Disputes with Violent

Groups.” International Studies Quarterly 59 (3): 461–76.

Carter, David B. 2016. “Provocation and the Strategy of Terrorist and Guerrilla Attacks.”

International Organization 70 (1): 133–73.

Chen, Andrew H. and Thomas F. Siems. 2004. “The Effects of Terrorism on Global Capital

Markets.” European Journal of Political Economy 20 (2): 249–66.

Chenoweth, Erica. 2010. “Democratic Competition and Terrorist Activity.” Journal of Politics

72 (1): 16–30.

Chenoweth, Erica. 2013. “Terrorism and Democracy.” Annual Review of Political Science 16

76

(1): 355–78.

Chesney, Marc, Ganna Reshetar, and Mustafa Karaman. 2011. “The Impact of Terrorism on

Financial Markets: An Empirical Study.” Journal of Banking & Finance 35 (2): 253–67.

Conrad, Justin and Kevin Greene. 2015. “Competition, Differentiation, and the Severity of

Terrorist Attacks.” Journal of Politics 77 (2): 546–61.

Cornes, Richard and Todd Sandler. 1996. The Theory of Externalities, Public Goods, and Club

Goods, 2nd ed. Cambridge: Cambridge University Press.

Crenshaw, Martha. 1981. “The Causes of Terrorism.” Comparative Politics 13 (4): 379–99.

Cronin, Audrey K. 2006. “How al-Qaida Ends: The Decline and Demise of Terrorist Groups.”

International Security 31 (1): 7–48.

Das, Satya. 2008. “Some Mechanisms of Terror Cycles.” Journal of Economic Behavior and

Organization 67 (3–4): 644–56.

Davis, Darren W. and Brian D. Silver. 2004. “Civil Liberties vs. Security: Public Opinion in

the Context of the Terrorist Attacks on America.” American Journal of Political Science

48 (1): 38–46.

Daxecker, Ursula E. and Michael L. Hess. 2013. “Repression Hurts: Coercive Government

Responses and the Demise of Terrorist Campaigns.” British Journal of Political Science

43 (3): 559–77.

de la Calle, Luis and Ignacio Sánchez-Cuenca. 2012. “Rebels without a Territory: An Analysis

of Nonterritorial Conflicts in the World, 1970–1977.” Journal of Conflict Resolution 56

(4): 580–603.

Department of Homeland Security (DHS). 2016. Budget in Brief: Fiscal Year 2017. Available

at https: https://www.dhs.gov/sites/default/files/publications/FY2017BIB.pdf (Accessed

November 12, 2017).

77

De Sousa, José, Daniel Mirza, and Thierry Verdier. 2009. “Trade and the Spillovers of

Transnational Terrorism.” Swiss Journal of Economics and Statistics 145 (4): 453–61.

Dragu, Tiberiu. 2011. “Is There a Trade-Off between Security and Liberty? Executive Bias,

Privacy Protections, and Terrorism Prevention.” American Political Science Review 105

(1): 64–78.

Drakos, Konstantinos. 2004. “Terrorism-Induced Structural Shifts in Financial Risk: Airline

Stocks in the Aftermath of the September 11th Terror Attacks.” European Journal of

Political Economy 20 (2): 436–46.

Drakos, Konstantinos and Andreas Gofas. 2006. “In Search of the Average Transnational

Terrorist Attack Venue.” Defence and Peace Economics 17 (2): 73–93.

Drakos, Konstantinos and Ali M. Kutan. 2003. “Regional Effects of Terrorism on Tourism in

Three Mediterranean Countries.” Journal of Conflict Resolution 47 (5): 621–41.

Dreher, Axel. 2006. “Does Globalization Affect Growth? Empirical Evidence from a New

Index.” Applied Economics 38 (10): 1091–110.

Dreher, Axel and Justin A. V. Fischer. 2010. “Government Decentralization as a Disincentive

for Transnational Terror? An Empirical Analysis.” International Economic Review 51

(4): 981–1002.

Dreher, Axel and Andreas Fuchs. 2011. “Does Terror Increase Aid?” Public Choice 149 (3–4):

337–63.

Dugan, Laura and Erica Chenoweth. 2012. “Moving beyond Deterrence: The Effectiveness of

Raising the Expected Utility of Abstaining from Terrorism in Israel.” American

Sociological Review 77 (4): 597–624.

Eckstein, Zvi and Daniel Tsiddon. 2004. “Macroeconomic Consequences of Terror: Theory

and the Case of Israel.” Journal of Monetary Economics 51 (5): 971–1002.

78

Egger, Peter and Martin Gassebner. 2015. “International Terrorism as a Trade Impediment?”

Oxford Economic Papers 67 (1): 42–62.

Enders, Walter and Gary A. Hoover. 2012. “The Nonlinear Relationship between Terrorism and

Poverty.” American Economic Review 102 (3): 267–72.

Enders, Walter, Gary A. Hoover, and Todd Sandler. 2016. The Changing Nonlinear

Relationship between Income and Terrorism.” Journal of Conflict Resolution 60 (2):

195–225.

Enders, Walter and Paan Jindapon. 2010. “Network Externalities and the Structure of Terror

Networks.” Journal of Conflict Resolution 54 (2): 262–80.

Enders, Walter, Adolfo Sachsida, and Todd Sandler. 2006. “The Impact of Transnational

Terrorism on U.S. Foreign Direct Investment.” Political Research Quarterly 59 (4):

517–31.

Enders, Walter and Todd Sandler. 1993. “The Effectiveness of Anti-Terrorism Policies: A

Vector-Autoregression-Intervention Analysis.” American Political Science Review 87

(4): 829–44.

Enders, Walter and Todd Sandler. 1996. “Terrorism and Foreign Direct Investment in Spain

and Greece.” Kyklos 49 (3): 331–52.

Enders, Walter and Todd Sandler. 1999. “Transnational Terrorism in the Post-Cold War Era.”

International Studies Quarterly 43 (2): 145–67.

Enders, Walter and Todd Sandler. 2000. “Is Transnational Terrorism Becoming More

Threatening? A Time-Series Investigation.” Journal of Conflict Resolution 44 (3): 307–

32.

Enders, Walter and Todd Sandler. 2002. “Patterns of Transnational Terrorism, 1970–99:

Alternative Time Series Estimates.” International Studies Quarterly 46 (2): 145–65.

79

Enders, Walter and Todd Sandler. 2004. “What Do We Know about the Substitution Effect in

Transnational Terrorism.” In Researching Terrorism: Trends, Achievements, Failures,

edited by Andrew Silke, 119–37. Ilford, UK: Frank Cass.

Enders, Walter and Todd Sandler. 2005a. “After 9/11: Is It All Different Now?” Journal of

Conflict Resolution 49 (2): 259–77.

Enders, Walter and Todd Sandler. 2005b. “Transnational Terrorism 1968–2000: Thresholds,

Persistence, and Forecasts.” Southern Economic Journal 71 (3): 467–82.

Enders, Walter and Todd Sandler. 2006. “Distribution of Transnational Terrorism among

Countries by Income Classes and Geography after 9/11.” International Studies Quarterly

50 (2): 367–93.

Enders, Walter and Todd Sandler. 2011. “Who Adopts MIND/FIND in INTERPOL’s Fight

against International Crime and Terrorism?” Public Choice 149 (3–4): 263–80.

Enders, Walter and Todd Sandler. 2012. The Political Economy of Terrorism, 2nd ed. New

York: Cambridge University Press.

Enders, Walter, Todd Sandler, and Jon Cauley. 1990. “UN Conventions, Technology, and

Retaliation in the Fight against Terrorism: An Econometric Evaluation.” Terrorism and

Political Violence 2 (1): 83–105.

Enders, Walter, Todd Sandler, and Khusrav Gaibulloev. 2011. “Domestic Versus Transnational

Terrorism: Data, Decomposition, and Dynamics.” Journal of Peace Research 48 (3):

355–71.

Enders, Walter, Todd Sandler, and Gerald F. Parise. 1992. “An Econometric Analysis of the

Impact of Terrorism on Tourism.” Kyklos 45 (4): 531–54.

Enders, Walter and Xuejuan Su. 2007. “Rational Terrorists and Optimal Network Structure.”

Journal of Conflict Resolution 51 (1): 33–57.

80

Engene, Jan O. 2007. “Five Decades of Terrorism in Europe: The TWEED dataset.” Journal

of Peace Research 44 (1): 109–21.

Eubank, William L. and Leonard B. Weinberg. 1994. “Does Democracy Encourage

Terrorism?” Terrorism and Political Violence 6 (4): 417–35.

Eubank, William L. and Leonard B. Weinberg. 2001. “Terrorism and Democracy: Perpetrators

and Victims.” Terrorism and Political Violence 13 (1): 155–64.

Eyerman, Joe. 1998. “Terrorism and Democratic States: Soft Targets or Accessible Systems?”

International Interactions 24 (2): 151–70.

Fair, C. Christine. 2010. “Drone Wars.” Available at

http://foreignpolicy.com/2010/05/28/drone-wars-2 (Accessed on November 8, 2017).

Faria, João R. 2003. “Terror Cycles.” Studies in Nonlinear Dynamics and Econometrics 7 (1):

1–11.

Faria, João R. and Daniel G. Arce. 2005. “Terror Support and Recruitment.” Defence and

Peace Economics 16 (4): 263–73.

Fearon, James D. and David D. Laitin. 2003. “Ethnicity, Insurgency, and Civil War.” American

Political Science Review 97 (1): 75–90.

Feinstein, Jonathan S. and Edward H. Kaplan. 2010. “Analysis of a Strategic Terror

Organization.” Journal of Conflict Resolution 54 (2): 281–302.

Findley, Michael E. and Joseph K. Young. 2012. “Terrorism and Civil War: A Spatial and

Temporal Approach to a Conceptual Problem.” Perspectives on Politics 10 (2): 285–305.

Fleck, Robert K. and Christopher Kilby. 2010. “Changing Aid Regimes? US Foreign Aid from

the Cold War to War on Terror.” Journal of Development Economics 91 (1): 185–97.

Frey, Bruno S. and Simon Luechinger. 2003. “How to Fight Terrorism: Alternatives to

Deterrence.” Defence and Peace Economics 14 (4): 237–49.

81

Frey, Bruno S., Simon Luechinger, and Alois Stutzer. 2009. “The Life Satisfaction Approach to

Valuing Public Goods: The Case of Terrorism.” Public Choice 138 (3–4): 317–45.

Gaibulloev, Khusrav. 2015. “Terrorist Group Location Decision: An Empirical Investigation.”

Oxford Economic Papers 67 (1): 21–41.

Gaibulloev, Khusrav, James A. Piazza, and Todd Sandler. 2017. “Regime Types and

Terrorism.” International Organization 71 (3), 491–522.

Gaibulloev, Khusrav and Todd Sandler. 2008. “Growth Consequences of Terrorism in Western

Europe.” Kyklos 61 (3): 411–24.

Gaibulloev, Khusrav and Todd Sandler. 2009a. “Hostage Taking: Determinants of Terrorist

Logistical and Negotiation Success.” Journal of Peace Research 46 (6): 739–56.

Gaibulloev, Khusrav and Todd Sandler. 2009b. “The Impact of Terrorism and Conflicts on

Growth in Asia.” Economics & Politics 21 (3): 359–83.

Gaibulloev, Khusrav and Todd Sandler. 2011. “The Adverse Effect of Transnational and

Domestic Terrorism on Growth in Africa.” Journal of Peace Research 48 (3): 355–71.

Gaibulloev, Khusrav and Todd Sandler. 2013. “Determinants of the Demise of Terrorist

Organizations.” Southern Economic Journal 79 (4): 774–92.

Gaibulloev, Khusrav and Todd Sandler. 2014. “An Empirical Analysis of Alternative Ways

That Terrorist Groups End.” Public Choice 160 (1–2): 25–44.

Gaibulloev, Khusrav, Todd Sandler, and Charlinda Santifort. 2012. “Assessing the Evolving

Threat of Terrorism.” Global Policy 3 (2): 135–144.

Gaibulloev, Khusrav, Todd Sandler, and Donggyu Sul. 2013. “Common Drivers of

Transnational Terrorism: Principal Component Analysis.” Economic Inquiry 51 (1):

707–21.

Gaibulloev, Khusrav, Todd Sandler, and Donggyu Sul. 2014. “Dynamic Panel Analysis under

82

Cross-Sectional Dependence.” Political Analysis 22 (2): 258–73.

Gardeazabal, Javier and Todd Sandler. 2015. “INTERPOL’s Surveillance Network in Curbing

Transnational Terrorism.” Journal of Policy Analysis and Management 34 (4): 761–80.

Gassebner, Martin, Richard Jong-A-Pin, and Jochen O. Mierau. 2011. “Terroism and Cabinet

Duration.” International Economic Review 52 (4): 1253–70.

Gassebner, Martin and Simon Luechinger. 2011. “Lock, Stock, and Barrel: A Comprehensive

Assessment of the Determinants of Terror.” Public Choice 149 (3–4): 235–61.

George, Justin. 2018. “State Failure and Transnational Terrorism: An Empirical Analysis.”

Journal of Conflict Resolution 62 (3): 471–95.

Gould, Eric D. and Esteban F. Klor. 2010. “Does Terrorism Work?” Quarterly Journal of

Economics 125 (4): 1459–510.

Hegghammer, Thomas. 2013. “Should I Stay or Should I Go? Explaining Variation in Western

Jihadists’ Choice between Domestic and Foreign Fighting.” American Political Science

Review 107 (1): 1–15.

Hoffman, Bruce. 2006. Inside Terrorism, Revised and Expanded Edition. New York:

Columbia University Press.

Iannaccone, Laurence R. 1992. “Sacrifice and Stigma: Reducing Free-Riding in Cults,

Communes, and Other Collectives.” Journal of Political Economy 100 (2): 271–91.

Iannaccone, Laurence R. and Eli Berman. 2006. “Religious Extremism: The Good, the Bad,

and the Deadly.” Public Choice 128 (1–2): 109–29.

Ito, Harumi and Darin Lee. 2005. “Assessing the Impact of the September 11 Terrorist Attacks

on U.S. Airline Demand.” Journal of Economics and Business 57 (1): 79–95.

Jacobson, Daniel and Edward H. Kaplan. 2007. “Suicide Bombings and Targeted Killings in

(Counter-) Terror Games.” Journal of Conflict Resolution 51 (5): 772–92.

83

Jaeger, David A., Esteban F. Klor, Sami H. Miaari, and M. Daniele Paserman. 2012. “The

Struggle for Palestinian Hearts and Minds: Violence and Public Opinion in the Second

Intifada.” Journal of Public Economics 96 (3–4): 354–68.

Jaeger, David A., Esteban F. Klor, Sami H. Miaari, and M. Daniele Paserman. 2015. “Can

Militants Use Violence to Win Public Support? Evidence from the Second Intifada.”

Journal of Conflict Resolution 59 (3): 528–49.

Jaeger, David A. and M. Daniele Paserman. 2008. “The Cycle of Violence? An Empirical

Analysis of Fatalities in the Palestinian-Israeli Conflict.” American Economic Review 98

(4): 1591–1604.

Jaeger, David A. and M. Daniele Paserman. 2009. “The Shape of Things to Come? On the

Dynamics of Suicide Attacks and Targeted Killings.” Quarterly Journal of Political

Science 4 (4): 315–42.

Jaeger, David A. and Zahra Siddique. 2011. “Are Drone Strikes Effective in Afghanistan and

Pakistan? On the Dynamics of Violence between and United States and the Taliban.”

Discussion Paper 6262, The Institute for the Study of Labor (IZA), Bonn, Germany.

Jindapon, Paan and William S. Neilson. 2009. “The Impact of Societal Risk Attitudes on

Terrorism and Counterterrorism.” Economics & Politics 21 (3): 433–51.

Johnston, Patrick B. and Anoop K. Sarbahi. 2016. “The Impact of US Drone Strikes on

Terrorism in Pakistan.” International Studies Quarterly 60 (2): 203–19.

Jones, Seth G. and Martin C. Libicki. 2008. How Terrorist Groups End: Lessons for Countering

al Qa’ida. Santa Monica, CA: RAND.

Kalyvas, Stathis N. and Laia Balcells. 2010. “International System and Technologies of

Rebellion: How the End of the Cold War Shaped Internal Conflict.” American Political

Science Review 104 (3): 415–29.

84

Krueger, Alan B. 2007. What Makes a Terrorist: Economics and the Roots of Terrorism.

Princeton, NJ: Princeton University Press.

Krueger, Alan B. 2008. “What Makes a Homegrown Terrorist? Human Capital and

Participation in Domestic Islamic Terrorist Groups in the USA.” Economics Letters 101

(3): 293–96.

Krueger, Alan B. and David D. Laitin. 2008. “Kto Kogo? A Cross-Country Study of the

Origins and Targets of Terrorism.” In Terrorism, Economic Development and Political

Openness, edited by Philip Keefer and Norman Loayza, 143–73. New York: Cambridge

University Press.

Krueger, Alan B. and Jitka Malečková. 2003. “Education, Poverty, and Terrorism: Is There a

Causal Connection?” Journal of Economic Perspectives 17 (4): 119–44.

Kunreuther, Howard and Geoffrey Heal. 2003. “Interdependent Security.” Journal of Risk and

Uncertainty 26 (2–3): 231–49.

Kunreuther, Howard, Erwann Michel-Kerjan, and Beverly Porter. 2003. “Assessing, Managing

and Financing Extreme Events: Dealing with Terrorism.” Working Paper 10179,

National Bureau of Economic Research, Cambridge, MA.

Kurrild-Klitgaard, Peter, Morgens K. Justesen, and Robert Klemmensen. 2006. “The Political

Economy of Freedom, Democracy, and Transnational Terrorism.” Public Choice 128 (1–

2): 289–315.

Kydd, Andrew and Barbara F. Walter. 2002. “Sabotaging the Peace: The Politics of Extremist

Violence.” International Organization 56 (2): 263–96.

Lai, Brian. 2007. “Draining the Swamp: An Empirical Examination of the Production of

International Terrorism.” Conflict Management and Peace Science 24 (4): 297–310.

Laitin, David D. and Jacob N. Shapiro. 2008. “The Political, Economic, and Organizational

85

Sources of Terrorism.” In Terrorism, Economic Development and Political Openness,

edited by Philip Keefer and Norman Loayza, 209–32. New York: Cambridge University

Press.

Landes, William M. 1978. “An Economic Study of US Aircraft Hijackings, 1961–1976.”

Journal of Law and Economics 21 (1): 1–31.

Lapan, Harvey E. and Todd Sandler. 1988. “To Bargain or Not to Bargain: That Is the

Question.” American Economic Review 78 (2): 16–20.

Lapan, Harvey E. and Todd Sandler. 1993. “Terrorism and Signalling.” European Journal of

Political Economy 9 (3): 383–97.

Li, Quan. 2005. “Does Democracy Promote Transnational Terrorist Incidents?” Journal of

Conflict Resolution 49 (2): 278–97.

Li, Quan and Drew Schaub. 2004. “Economic Globalization and Transnational Terrorism: A

Pooled Time-Series Analysis.” Journal of Conflict Resolution 48 (2): 230–58.

Memon, Nasrullah, Jonathan D. Farley, David L. Hicks, and Torben Rosenørn (eds.). 2009.

Mathematical Methods in Counterterrorism. Wien: Springer-Verlag.

Merari, Ariel. 1993. “Terrorism as a Strategy of Insurgency.” Terrorism and Political Violence

5 (4): 213–51.

Merari, Ariel. 2004. “Suicide Terrorism in the Context of the Israeli-Palestinian Conflict.”

Presentation at the Suicide Terrorism Research Conference, October 25–26, National

Institute of Justice, Washington, DC.

Mickolus, Edward F., Todd Sandler, and Jean M. Murdock. 1989. International Terrorism in

the 1980s: A Chronology of Events, 2 vols. Ames, IA: Iowa State University.

Mickolus, Edward F., Todd Sandler, Jean M. Murdock, and Peter Flemming. 2017.

International Terrorism: Attributes of Terrorist Events, 1968–2016 (ITERATE). Pointe

86

Vedra, FL: Vinyard Software.

Mirza, Daniel and Thierry Verdier. 2014. “Are Lives a Substitute for Livelihoods? Terrorism,

Security, and U.S. Bilateral Imports.” Journal of Conflict Resolution 58 (6): 943–75.

National Consortium for the Study of Terrorism and Responses to Terrorism (START). 2017.

Global Terrorism Database (GTD), CD-ROM. College Park, MD: University of

Maryland. Available at www.start.umd.edu/gtd/ (Accessed November 12, 2017).

Nemeth, Stephen. 2014. “The Effect of Competition on Terrorist Group Operations. Journal of

Conflict Resolution 58 (2): 336–62.

Neumayer, Eric. 2004. “The Impact of Political Violence on Tourism: Dynamic Cross-

National Estimation.” Journal of Conflict Resolution 48 (2): 259–81.

Neumayer, Eric and Thomas Plümper. 2016. “Spatial Spill-overs from Terrorism on Tourism:

Western Victims in Islamic Destination Countries.” Public Choice 169 (3-4): 195–206.

Nitsch, Volker and Dieter Schumacher. 2004. “Terrorism and International Trade: An

Empirical Investigation.” European Journal of Political Economy 20 (2): 423–33.

Overgaard, Per Baltzer. 1994. “The Scale of Terrorist Attacks as a Signal of Resources.”

Journal of Conflict Resolution 38 (3): 452–78.

Pape, Robert A. 2003. “The Strategic Logic of Suicide Terrorism.” American Political Science

Review 97 (3): 343–61.

Pape, Robert A. 2005. Dying to Win: The Strategic Logic of Suicide Terrorism. New York:

Random House.

Phillips, Brian J. 2014. “Terrorist Group Cooperation and Longevity.” International Studies

Quarterly 58 (2): 336–47.

Phillips, Brian J. 2015. “What Is a Terrorist Group? Conceptual Issues and Empirical

Implications.” Terrorism and Political Violence 27 (2): 225–42.

87

Piazza, James A. 2008. “Incubators of Terror: Do Failed and Failing States Promote

Transnational Terrorism?” International Studies Quarterly 52 (3): 469–88.

Piazza, James A. 2011. “Poverty, Minority Economic Discrimination and Domestic Terrorism.”

Journal of Peace Research 48 (3): 339–53.

Pizam, Abraham and Aliza Fleischer. 2002. “Severity versus Frequency of Acts of Terrorism:

Which Has a Larger Impact on Tourism Demand?” Journal of Travel Research 40 (3):

337–9.

Powell, Robert. 2007. “Defending against Terrorist Attacks with Limited Resources.”

American Political Science Review 101 (3): 527–41.

RAND. 2012. Database of Worldwide Terrorism Incidents, available at

http://www.rand.org/nsrd/projects/terrorism-incidents.html (Assessed August 12, 2012).

Rapoport, David C. 2004. “Modern Terror: The Four Waves.” In Attacking Terrorism:

Elements of a Grand Strategy, edited by Audrey Kurth Cronin and James M. Ludes, 46–

73. Washington, DC: Georgetown University Press.

Rohner, Dominic and Bruno S. Frey. 2007. “Blood and Ink! The Common-Interest Game

between Terrorists and the Media.” Public Choice 133 (1–2): 129–45.

Rosendorff, B. Peter and Todd Sandler. 2004. “Too Much of a Good Thing? The Proactive

Response Dilemma.” Journal of Conflict Resolution 48 (5): 657–71.

Rosendorff, B. Peter and Todd Sandler. 2010. “Suicide Terrorism and the Backlash Effect.”

Defence and Peace Economics 21 (5–6): 443–57.

Sageman, Marc. 2004. Understanding Terror Networks. Philadelphia, PA: University of

Pennsylvania Press.

Sageman, Marc. 2008. Leaderless Jihad: Terror Networks in the Twenty-First Century.

Philadelphia, PA: University of Pennsylvania Press.

88

Sambanis, Nicholas. 2008. “Terrorism and Civil War.” In Terrorism, Economic Development

and Political Openness, edited by Philip Keefer and Norman Loayza, 174–206. New

York: Cambridge University Press.

San-Akca, Belgin. 2014. “Democracy and Vulnerability: An Exploitation Theory of

Democracies by Terrorists.” Journal of Conflict Resolution 58 (7): 1285–310.

Sandler, Todd. 1995. “On the Relationship between Democracy and Terrorism.” Terrorism and

Political Violence 7 (4): 1–9.

Sandler, Todd. 2003. “Collective Action and Transnational Terrorism.” World Economy 26 (6):

779–802.

Sandler, Todd. 2005. “Collective versus Unilateral Responses to Terrorism.” Public Choice

124 (1–2): 75–93.

Sandler, Todd. 2010. “Terrorism Shocks: Domestic versus Transnational Responses.” Studies

in Conflict & Terrorism 33 (10): 893–910.

Sandler, Todd and Daniel G. Arce. 2003. “Terrorism & Game Theory.” Simulation & Gaming

34 (3): 319–37.

Sandler, Todd, Daniel G. Arce, and Walter Enders. 2009. “Transnational Terrorism.” In Global

Crises, Global Solutions, 2nd ed., edited by Bjørn Lomborg, 516–62. Cambridge:

Cambridge University Press.

Sandler, Todd, Daniel G. Arce, and Walter Enders. 2011. “An Evaluation of INTERPOL’s

Cooperative-Based Counterterrorism Linkage.” Journal of Law and Economics 54 (1):

79–110.

Sandler, Todd and Keith Hartley. 1995. The Economics of Defense. Cambridge: Cambridge

University Press.

Sandler, Todd and Harvey E. Lapan. 1988. “The Calculus of Dissent: An Analysis of

89

Terrorists’ Choice of Targets.” Synthese 76 (2): 245–61.

Sandler, Todd and Kevin Siqueira. 2006. “Global Terrorism: Deterrence versus Preemption.”

Canadian Journal of Economics 50 (4): 1370–87.

Sandler, Todd and Kevin Siqueira. 2009. “Games and Terrorism: Recent Developments.”

Simulation & Gaming 40 (2): 164–92.

Sandler, Todd, John Tschirhart, and Jon Cauley. 1983. “A Theoretical Analysis of

Transnational Terrorism.” American Political Science Review 77 (1): 36–54.

Santifort, Charlinda and Todd Sandler. 2013. “Terrorist Success in Hostage-Taking Missions:

1978–2010.” Public Choice 156 (1–2): 125–37.

Santifort-Jordan, Charlinda and Todd Sandler. 2014. “An Empirical Study of Suicide

Terrorism: A Global Analysis.” Southern Economic Journal 80 (4): 981–1001.

Savun, Burcu and Brian J. Phillips. 2009. “Democracy, Foreign Policy, and Terrorism.”

Journal of Conflict Resolution 53 (6): 878–904.

Schneider, Friedrich, Tilman Brück, and Daniel Meierrieks. 2015. “The Economics of

Counterterrorism: A Survey.” Journal of Economic Surveys 29 (1): 131–57.

Selten, Reinhard. 1977. “A Simple Game Model of Kidnappings.” In Mathematical Economics

and Game Theory: Essays in Honor of Oskar Morgenstern, edited by R. Henn and O.

Moeschlin, 139–55. Berlin: Springer-Verlag.

Shapiro, Jacob N. 2007. “Terrorist Organizations’ Vulnerabilities and Inefficiencies.” In

Terrorism Financing and State Responses: A Comparative Perspective, edited by Jeanne

K. Giraldo and Harold A. Trinkunas, 308–13. Stanford, CA: Stanford University Press.

Shapiro, Jacob N. 2013. The Terrorist’s Dilemma: Managing Violent Covert Organizations.

Princeton, NJ: Princeton University Press.

Shapiro, Jacob N. and David A. Siegel. 2007. “Underfunding in Terrorist Organizations.”

90

International Studies Quarterly 51 (2): 405–29.

Siqueira, Kevin. 2005. “Political and Militant Wings within Dissident Movements and

Organizations.” Journal of Conflict Resolution 49 (2): 218–36.

Siqueira, Kevin and Todd Sandler. 2006. “Terrorists versus the Government: Strategic

Interaction, Support, and Sponsorship.” Journal of Conflict Resolution 50 (6): 878–98.

Siqueira, Kevin and Todd Sandler. 2007. “Terrorist Backlash, Terrorism Mitigation, and Policy

Delegation.” Journal of Public Economics 91 (9): 1800–15.

Siqueira, Kevin and Todd Sandler. 2010. “Terrorist Networks, Support, and Delegation.”

Public Choice 142 (1–2): 237–53.

Sloboda, Brian W. 2003. “Assessing the Effects of Terrorism on Tourism by the Use of Time

Series Methods.” Tourism Economics 9 (2): 179–90.

Tavares, Jose. 2004. “The Open Society Assesses its Enemies: Shocks, Disasters and Terrorist

Attacks.” Journal of Monetary Economics 51 (5): 1039–70.

Viscusi, W. Kip and Richard J. Zeckhauser. 2003. “Sacrificing Civil Liberties to Reduce

Terrorism Risks.” Journal of Risk and Uncertainty 26 (2–3): 99–120.

Wilkinson, Paul. 2001. Terrorism versus Democracy: The Liberal State Response. London:

Frank Cass.

Wintrobe, Ronald. 2006. “Extremism, Suicide Terror, and Authoritarianism.” Public Choice

128 (1): 169–95.

Young, Joseph K. and Michael G. Findley. 2011. “Can Peace Be Purchased? A Sectoral-Level

Analysis of Aid’s Influence on Transnational Terrorism.” Public Choice 149 (3–4): 365–

81.

Zhuang, Jun and Vicki M. Bier. 2007. “Balancing Terrorism and Natural Disasters – Defensive

Strategy with Endogenous Attacker Effort.” Operations Research 55 (5): 976–91.

91

Zussman, Asaf and Noam Zussman. 2006. “Assassinations: Evaluating the Effectiveness of an

Israeli Counterterrorism Policy Using Stock Market Data.” Journal of Economic

Perspectives 20 (2): 193–206.

TABLE 1

DISTRIBUTION OF TERRORIST GROUPS BY IDEOLOGY AND BASE COUNTRY OF OPERATION’S CHARACTERISTICS: BEFORE AND AFTER 2001

Ideology Average Number Number

Base country’s economy Base country’s regime type

Low IncomeMiddle Income High Income Not Free Partly Free Free

Groups started before 2001

Left Wing 5.3 198 (38) 26 (21) 102 (43) 69 (44) 21 (20) 64 (39) 113 (45)

Nationalist 4.9 192 (37) 56 (46) 78 (33) 58 (37) 44 (42) 60 (37) 88 (35)

Religious 2.6 90 (17) 35 (29) 47 (20) 8 (5) 36 (34) 34 (21) 20 (8)

Right Wing 1.1 38 (7) 4 (3) 13 (5) 21 (13) 5 (5) 5 (3) 28 (11)

Total 13.9 518 (100) 121 (100) 240 (100) 156 (100) 106 (100) 163 (100) 249 (100)

Groups started during 2001–2006

Left Wing 6.0 36 (28) 2 (9) 18 (23) 16 (57) 0 (0) 13 (35) 23 (52)

Nationalist 6.3 38 (29) 6 (27) 25 (31) 7 (25) 11 (22) 13 (35) 14 (32)

Religious 8.7 52 (40) 14 (64) 34 (43) 4 (14) 38 (78) 9 (24) 5 (11)

Right Wing 0.7 4 (3) 0 (0) 3 (4) 1 (4) 0 (0) 2 (5) 2 (5)

Total 21.7 130 (100) 22 (100) 80 (100) 28 (100) 49 (100) 37 (100) 44 (100) Notes: All data come from Jones and Libicki (2008). Numbers in parentheses are percentages; e.g., out of 518 groups started before 2001, 38% are left wing. The average annual number of groups for pre-2001 period is computed using number of groups that started during 1968–2001 and dividing by 34. Therefore, we exclude 47 groups that started sometime before 1968 in the average column calculation.

TABLE 2 REGIONAL DISTRIBUTION OF TERRORIST GROUPS BY IDEOLOGY: BEFORE AND AFTER 2002

Ideology EAP ECA LAC MENA NA SA SSA Total

Groups started before 2002

Left Wing 13 73 75 12 10 6 10 199

Nationalist 12 69 6 49 11 38 21 206

Religious 16 10 1 34 4 24 7 96

Right Wing 2 9 10 0 2 0 3 26

Total 43 161 92 95 27 68 41 527

Groups started during 2002–2006

Left Wing 1 12 13 0 1 2 0 29

Nationalist 2 12 0 12 0 5 1 32

Religious 0 3 0 28 0 10 0 41

Right Wing 1 0 1 0 0 0 1 3

Total 4 27 14 40 1 17 2 105 Notes: Terrorist group data come from Jones and Libicki (2008). Regions are defined following World Bank classification: South Asia, SA; Europe and Central Asia, ECA; Middle East and North Africa, MENA; Latin America and the Caribbean, LAC; East Asia and Pacific, EAP; sub-Saharan Africa, SSA; and North America, NA. There are 586 terrorist groups but the total number of observations is 633; some groups have multiple bases of operation in different regions.

TABLE 3 DISTRIBUTION OF TERRORIST INCIDENTS BY GROUP IDEOLOGY: 2002–2007

Ideology Number of

groups

Number of groups ended by

2007 Number of incidents

Number of casualties

Number of incidents per

group

Number of casualties per

group

Left Wing 29 13 322 919 11.1 31.7

Nationalist 30 13 124 2415 4.1 80.5

Religious 41 12 216 3482 5.3 84.9

Right Wing 3 2 12 11 4.0 3.7

Total 103 40 674 6827 6.5 68.3 Notes: Terrorist group data come from Jones and Libicki (2008). Data on terrorist (both transnational and domestic) incidents come from RAND (2012).

TABLE 4 DISTRIBUTION OF TRANSNATIONAL INCIDENTS AND CASUALTIES BY TYPE OF

ATTACKS

Bombing Hostage Armed attack Assassination Others Total

Number of incidents

1968–2001 5506 1772 1462 1049 2388 12177

2002–2016 849 590 628 197 213 2477

Average number of incidents per year

1968–2001 161.9 52.1 43.0 30.9 70.2 358

2002–2016 56.6 39.3 41.9 13.1 14.2 165

Number of casualties

1968–2001 21103 3581 5600 1877 387 32548

2002–2016 18367 2464 5495 849 197 27372

Average number of casualties per year

1968–2001 620.7 105.3 164.7 55.2 11.4 957

2002–2016 1224.5 164.3 366.3 56.6 13.1 1825

TABLE 5

DISTRIBUTION OF DOMESTIC INCIDENTS AND CASUALTIES BY TYPE OF ATTACKS

Bombing Hostage Armed attack Assassination Others Total

Number of incidents

1970–2001 19942 210 8273 7365 4769 40559

2002–2016 31715 4326 13342 4558 3737 57678

Average number of incidents per year

1970–2001 623.2 6.6 258.5 230.2 149.0 1267

2002–2016 2114.3 288.4 889.5 303.9 249.1 3845

Number of casualties

1970–2001 79728.0 503 50262 13092 12413.0 155998

2002–2016 255581.9 17530.5 69959.3 11830.2 6803.8 361706

Average number of casualties per year

1970–2001 2491.5 15.7 1570.7 409.1 387.9 4875

2002–2016 17038.8 1168.7 4664.0 788.7 453.6 24114

Figure 1. Number of Transnational and Domestic Terrorist Incidents

0

2000

4000

6000

8000

10000

12000

14000

0

100

200

300

400

500

600

700

1968 1972 1976 1980 1984 1988 1992 1996 2000 2004 2008 2012 2016

Nu

mb

er o

f d

omes

tic

inci

den

ts

Nu

mb

er o

f tr

ansn

atio

nal

inci

den

tsTransnational Domestic

Figure 2. Number of Casualties per Incident

0.0

2.0

4.0

6.0

8.0

10.0

12.0

14.0

16.0

18.0

20.0

1968 1972 1976 1980 1984 1988 1992 1996 2000 2004 2008 2012 2016

Cas

ual

ties

per

inci

den

tTransnational Domestic

Figure 3. Proportion of Transnational and Domestic Terrorist Events with Casualties

0

10

20

30

40

50

60

70

80

90

1968 1972 1976 1980 1984 1988 1992 1996 2000 2004 2008 2012 2016

Per

cen

tage

Transnational Domestic

Figure 4. Number and Percentage of Transnational Attacks against US Interests per Year

0.0

10.0

20.0

30.0

40.0

50.0

60.0

70.0

80.0

90.0

0

50

100

150

200

250

300

1968 1972 1976 1980 1984 1988 1992 1996 2000 2004 2008 2012 2016

Per

cen

tage

Nu

mb

er o

f in

cid

ents

Number of incidents Percentage of incidents

Figure 5. Distribution of Transnational Incidents by Type of Attacks for 1968–2001 and 2002–2016

0.00

5.00

10.00

15.00

20.00

25.00

30.00

35.00

40.00

45.00

50.00

Bombing Hostage Armed Attack Assassination

Per

cen

tage

1968-2001

2002-2016

Figure 6. Distribution of Domestic Incidents by Type of Attacks for 1970–2001 and 2002–2016

0.00

10.00

20.00

30.00

40.00

50.00

60.00

Bombing Hostage Armed Attack Assassination

Per

cen

tage

1970-2001

2002-2016

Figure 7a. Distribution of Transnational Incidents by Regions for 1968–2001

Figure 7b. Distribution of Transnational Incidents by Regions for 2002–2016

EAP8%

ECA37%

LAC21%

MENA17%

NA6%

SA4% SSA

7%

1968-2001

EAP5%

ECA13%

LAC3%

MENA35%

NA1%

SA20%

SSA23%

2002-2016

Figure 8a. Distribution of Domestic Incidents by Regions for 1970–2001

Figure 8b. Distribution of Domestic Incidents by Regions for 2002–2016

EAP6%

ECA21%

LAC42%

MENA9%

NA2%

SA13%

SSA7%

1970-2001

EAP8%

ECA6%

LAC2%

MENA37%

NA1%

SA35%

SSA11%

2002-2016

Figure 9a. Distribution of Transnational Incident Casualties by Regions for 1968–2001

Figure 9b. Distribution of Transnational Incident Casualties by Regions for 2002–2016

EAP8%

ECA30%

LAC10%

MENA31%

NA3%

SA8%

SSA10%

1968-2001

EAP8%

ECA18%

LAC1%

MENA38%

NA1%

SA17%

SSA17%

2002-2016

Figure 10a. Distribution of Domestic Incident Casualties by Regions for 1970–2001

Figure 10b. Distribution of Domestic Incident Casualties by Regions for 2002–2016

EAP8%

ECA12%

LAC21%

MENA16%

NA1%

SA29%

SSA13%

1970-2001

EAP4%

ECA4%

LAC1%

MENA50%

NA0%

SA28%

SSA13%

2002-2016

Figure 11. Cumulative Number of Transnational Terrorist Attacks by Target Type

1980

1999

0

500

1000

1500

2000

2500

3000

3500

4000

4500

5000

1968 1972 1976 1980 1984 1988 1992 1996 2000 2004 2008 2012 2016

Cu

mu

lati

ve n

um

ber

of

inci

den

ts

Private Parties

Officials

Business

Military